Evaluation Report

Warning This Web page has been archived on the Web.

Correctional Service Canada's Correctional Programs

Evaluation Branch
Performance Assurance Sector
January 2009

PDF

Acknowledgements

The evaluation team wishes to thank all those staff members throughout Correctional Service Canada (CSC) who daily collect and record crucial information on virtually all aspects of federal offender correctional programs. Without their diligence in accumulating this information throughout their daily routines, this evaluation would not be possible.

The evaluation team would like to express their appreciation to the members of the Correctional Programs Evaluation consultative group and executive steering committee, for their comments and feedback regarding the evaluation process, as well as peer reviewing the evaluation plan, the intreview questions, and the evaluation report. We also wish to thank the Research Branch, CSC, for their collaboration and contributions to the report; specifically, Dan Kunic for collaborating in the analysis and methodological approach, and Christa Gillis, Shauna Botts and Tammy Cabana for their research and written sections on the relevancy and continued relevancy and continued relevancy sections of the report. The evaluation team sincerely appreciates the participation of staff members in all regions for their valuable assistance during the evaluation, in facilitating the identification of key stakeholders, setting aside their time to contribute their opinions and experiences regarding correctional programs, and ensuring timely access to other staff members, and relevant documentation.

The evaluation team also wishes to thank all who facilitated the creation of this report. Specifically, Nicole Crutcher from the Department of Justice for updating the team on potential legislation impacting the demographic of the federal offender profile; Tim Thompson for his explanations and guidance regarding some of the intricacies of the Offender Management System. Many people within and outside the Evaluation Branch rallied together to provide assistance performing the key informant interviews; Kelly Parnett, Valerie Campbell, Vanessa Anastasopoulos, Kelly Taylor, Megan Rooney, and Amanda Nolan, for their contribution to the qualitative data collection; Lynn Stewart, Jill Flight, Shauna Bottos, Linda Price, Anthony Coker, Drew Kingston, Phil Chitty, Denis Barbe, Sebastien Girard, and Phil Chitty provided for all their assistance in accessing and interpreting the pre–/post-program assessment battery data, and Annie Yessine for assisting in the final edits of the report.

EVALUATION PREPARED BY:

Mark Nafekh
Director, Evaluation

Nicole Allegri
Senior Evaluator

Yvonne Stys
Evaluation Officer

Tamara Jensen
Evaluation Analyst

EVALUATION TEAM MEMBERS

Angela Fabisiak
Senior Evaluator

Dennis Batten
Evaluation Officer

Michael Henighan
Evaluation Assistant

Elizabeth Loree
Evaluation Assistant

Hongping Li
Evaluation Officer

Meghan Chappell
Evaluation Assistant

Cara Scarfone
Evaluation Assistant

Correctional Service Canada’s
Correctional Program Evaluation

SIGNATURES

____________________
Lynn Garrow
Associate Assistant Commissioner
Policy and Research

____________________
Date

____________________
Pamela Yates
Director General
Evaluation Branch

____________________
Date

Executive Summary

Introduction

Correctional Service Canada1 (CSC) delivers structured interventions that address factors directly linked to offenders’ criminal behaviour (Correctional Service Canada [CSC], 2003). The primary goal of these correctional programs is to help offenders safely and successfully reintegrate into the community (Corrections and Conditional Release Act [CCRA], 1992) thereby enhancing public safety. The programs cover diverse content areas and include violence prevention programs, family violence programs, substance abuse interventions, sex offender programs, and life skills interventions2. Maintenance programs are also offered for several of these areas in order to ensure continuity of care to maximize treatment outcome. In addition, programs addressing the unique and diverse needs characteristic of Aboriginal and women offenders have been developed and implemented, with the aim of improving outcomes for these groups across several program areas.

The present report is an evaluation of CSC’s: (a) violent offender programs; (b) family violence programs; (c) substance abuse interventions; (d) sex offender programs; (e) life skills interventions; (f) community maintenance programs. The report measures achievements and outcomes as outlined in the Terms of Reference (2007). The report is summative in nature even though it incorporates aspects of both the formative and summative3 approaches towards evaluation. The report includes findings and recommendations regarding the implementation of the target programs; however, not all aspects of this area were examined in detail as they would be in a purely formative evaluation. The findings and recommendations contained in this report are designed to guide decisions regarding the design and delivery of the targeted correctional programs.

Evaluation Context

This evaluation of correctional programs was conducted between July 2007 and June 2008, and coincides with the evaluation of CSC’s employment strategy. Both evaluations are large in scope and correspond to two significant departmental reviews: (a) an independent panel review4 , a strategic review; and (b) an audit being conducted by the Office of the Auditor General (OAG) Canada.

Results of the first review, A Roadmap to Strengthening Public Safety (CSC Review Panel, 2007), were published in December 2007. The report was prepared by an independent Review Panel established by the Minister of Public Safety, the Honourable Stockwell Day. The review focussed on the operations, policies and plans of CSC in order to strengthen its contribution to public safety. Briefly, the report sets out a vision for transforming the federal corrections system based on an increased emphasis on offender responsibility and accountability; the elimination of drugs from prisons; the provision of more employment and employability skills for offenders, and the modernization of physical infrastructure.

The second review, as part of the renewal of Treasury Board of Canada Secretariat’s (TBS) Expenditure Management System (EMS), is being led by the TBS and is a review of departments’ program spending over a four-year cycle. This review will assess whether departments are achieving intended results, managing programs efficiently, and aligning programs with the government’s priorities. The strategic review of CSC’s program spending began in the fall of 2007 and will examine program effectiveness and opportunities for savings or reallocation.

Finally, in June 2007 the OAG began an audit of the efficiency and economy of CSC services. The audit proposed to examine the cost of various services and how they are: (a) structured (national, regional, institutional, community); (b) developed; (c) delivered; (d) evaluated; and (e) paid (FTEs, contract, contribution, etc.). Differences will be disaggregated in the audit across regions and federal institutions, as well as offender groups. The services which are expected to be a focus of the audit are: (a) health services (meeting offenders physical and mental health needs); (b) institutional services (meals, clean working and living environments, necessary supplies and clothing, options to purchase personal items); (c) accommodation services (suitable accommodation and living conditions); (d) security of offenders (security needs of offenders are identified, addressed within human rights context); (e) program development and delivery; (f) case management; and (g) CORCAN, as well as other relevant services.

Evaluation Strategy

The evaluation was conducted by the Evaluation Branch, CSC, and various components were reviewed by CSC’s Reintegration Programs and Research Branches. CSC’s Research Branch also contributed significantly towards the methodology used in the effectiveness analyses utilized in this report, as well as the literature review and components of the relevancy sections. This targeted evaluation was completed to inform decision-making on the future disposition of the programs and focused on success, cost–effectiveness, and continued relevancy.

The evaluation used both qualitative and quantitative methodologies. Information used to facilitate these analyses was collected through:

  • Interviews with key sources5;
  • Data derived from CSC’s automated data base – the Offender Management System (OMS), and the Canadian Police Information Centre (CPIC) data;
  • A review of relevant documentation (CSC National Program Guidelines, Standard Operating Practices relevant to correctional programming, Commissioner’s Directives (CDs) relevant to correctional programming, Corrections and Conditional Release Act (CCRA), detailed program descriptions provided by Reintegration Programs, information provided by CSC staff members to the evaluation team, and documents provided by offenders to the evaluation team6); and
  • A review of the relevant literature regarding the relevancy, success, and cost–effectiveness of correctional programs.

Interviews were conducted by the evaluation team and members of CSCs Research Branch, in person and by telephone during the months of April and May, 2008.

CSC’s OMS data were used to identify offenders program assignment status for a sample of first–term federal admissions between the years 1997 and 2007 (N = 48,601). Subsequently, two groups were created, an Intent–to–Treat (ITT) group of offenders who were exposed to the targeted correctional program through varying degrees of program participation, and a comparison group of offenders who were assigned to a program in which they did not participate. A multivariable approach to modeling was chosen to statistically control for any underlying differences that may be contributed to program effect other then program participation, such as offenders age, risk, motivation level, etc.7

Limitations

Given the broad scope of the evaluation, data analysis was limited primarily to databases housed within OMS that could easily be queried. Correctional plans, psychological reports and casework records were not examined but may have provided additional information.

Although the evaluation is comprehensive and was able to answer many meaningful questions, time constraints limited the scope of several aspects of the evaluation. For example, the gap between program referrals and actual enrolment could be investigated through qualitative interviews with staff members across the regions, and did not include input of offenders.

Many offenders have been identified as having multiple treatment needs and have participated in multiple programs. However, the inclusion of multiple dependent variables designating pre– post–participation in other correctional programs detracted from the true effect of the program being considered. A series of post–hoc analyses of multiple program participation was performed only for those programs which overlapped with the highest number of other programs; namely Living Skills and Substance Abuse programs. These analyses were not performed for other correctional programs. When a program was analysed for effectiveness against a comparison group, all results were attributed to participation in that program, regardless of other programs in which the group may have participated.

Program outcomes for women offenders may be limited for a variety of reasons. The Women Offenders Sex Offenders Programs are delivered on a one–on–one basis. The number of participants was therefore restricted, making it difficult to generate a large enough sample and comparison group for analyses. Similarly, the program structure for the Women Offenders Substance Abuse Programs did not facilitate analyses as offenders within the comparison group could have been exposed to components of other substance abuse programs. Finally, although the current report examined violence, it was limited to the Spirit of A Warrior Program. The Violence Prevention Program for Women Offenders is relatively new. Analyses were not conducted due to the insufficient number of participants having enrolled to date. It is anticipated that such analyses will be conducted once an adequate number of women offenders have taken part in the program.

It was assumed that all retrospective data were available to CSC at the time of program referral. Delays in receiving this information could have had an impact on the decision making process for program referrals. For example, new pending criminal charges at the time of the correctional plan development may not have been incorporated.

Correctional programs offered in the community and in institutions often engage community partners. Due to time constraints, the evaluation did not consult with volunteers of community groups on the functioning of correctional programs. As a mitigating strategy, these community partners will be engaged in a subsequent, more targeted evaluation on community reintegration initiatives.

Next, the continued relevancy section of the report demonstrates that positive correctional outcomes are reduced for particular demographics of the federal offender population. However, no baseline comparison was established to ascertain whether these outcomes are indeed above or below the expected level of outcomes within these particular groups.

Lastly, there may be a lack of precision in financial reporting for Aboriginal offender correctional programs potentially resulting from an overlap of activities and transfer of funding arrangements. This would explain, for instance, the relatively low expenditures on the In Search of Your Warrior Program.

Financial Expenditures

Actual expenditures for the STPI were $960,689.63 throughout the course of the Safer Tattooing Practices Pilot Initiative (7 months of start–up plus approximately 12 months implementation period). Expenditures reflect start–up costs (construction and equipment, supplies, staff expenditures on salaries, travel, and education and training costs, and technical contracts) and ongoing expenditures (see Table below).

Total Expenditures of Correctional Programs
   Fiscal Year 2005/2006
Expenditure
 Fiscal Year 2006/2007
Expenditure
 All Programs9 $53,232,479 $52,309,418
Correctional Programs10 $26,345,107 $25,947,480
 Sex Offender Programs11 $4,746,022 $4,391,955
 National Moderate Intensity Sex Offender $922,580 $618,471
 National High Intensity
National Low Intensity Sex Offender $170,852 $72,485
 Maintenance Sex Offender $307,870 $471,061
 Family Violence $2,330,671 $2,420,430
 High Intensity Family Violence $553,869 $575,032
 Maintenance $50,676 $1,816
Moderate Intensity Family Violence $506,926 $549,783
Aboriginal High Intensity Family Violence $0 $0
Living Skills Programs $4,489,750 $4,483,922
 Reasoning and Rehabilitation (discontinued this year) $2,494,164 $2,182,093
 Cognitive Maintenance $268,890 $192,700
 Anger and Emotions Management $1,226,090 $1,403,973
 Anger and Emotions Management Maintenance $30,639 $20,106
 Circles of Change $54 $36,537
 Aboriginal Basic Healing Program $189,784 $358,337
 Community Correctional Programs $1,359,201 $1,248,248
 Counter Point $1,184,936 $925,865
 Community Maintenance $174,265 $322,383
 Violent Offender Programs $3,866,253 $3,300,786
 Violence Prevention Program $1,546,764 $1,311,349
 In Search of Your Warrior – High Intensity $42,834 $116,751
 Spirit of A Warrior $523,604 $505,848
 Substance Abuse Programs $6,991,937 $7,499,484
 National Substance Abuse Program – High $1,717,724 $2,003,417
 National Substance Abuse Program – Moderate $2,349,381 $2,777,941
National Substance Abuse Program – Low $821,997 $591,348
National Substance Abuse Maintenance $773,657 $793,306
Women Offender Substance Abuse $784,690 $642,698
Aboriginal Offender Substance Abuse Program $75,911 $95,270
Women Offender Programs $622,771 $546,149
Women Offender Relapse Prevention $160,790 $109,757

Summary Findings

SUMMARY FINDING 1:

There exists a large body of evidence, both national and international, which confirms that targeting specific criminogenic needs is a relevant and effective means of enhancing public safety and reducing re–offending. As such, CSCs correctional programs are consistent with departmental and government–wide priorities.

SUMMARY FINDING 2:

The criteria used to refer offenders to specific correctional programs are appropriate, although supplementary assessment tools that are actuarial in nature and that incorporate static risk were shown to discriminate intensity level requirements more accurately.

SUMMARY FINDING 3:

For all programs with the exception of sex offender programs, there was a high proportion of offenders who met the referral criteria for a program who were not referred, and a proportion of offenders referred to a program who did not participate. There was also a proportion of offenders enrolled in programs for which they had not met the referral criteria. The large pool of candidates who met the referral criteria for a violence prevention program or high intensity substance abuse program was more likely to be serving shorter sentences in a maximum security institution. This group was also typically referred to the moderate intensity substance abuse programs.

SUMMARY FINDING 4:

Federal offenders who participated in CSC correctional programs that targeted an identified treatment need exhibited treatment–related change, and were more likely to be granted a discretionary release. In general, correctional program participation was not, however, associated with reductions in major institutional incidents. For community outcomes, program participation was associated with reductions in readmissions, including technical revocations and re–offending (with a non–violent, a violent and/or a sex offence in either federal or provincial jurisdictions). However these findings vary by program, intensity level, and demographic of the federal offender population. Community correctional outcomes were found to be more positive when there was a match between assessed program referral criteria and program intensity level. Conversely, positive effects were diminished when offenders participated in programs that did not correspond to their assessed need. Lastly, positive correctional outcomes were generally more pronounced for offenders who remained enrolled until the end of the program.

SUMMARY FINDING 5:

Correctional programs are delivered in federal institutions cost–efficiently, as between 70 to 90 cents of every dollar spent on programs is allocated towards successful completions. The community maintenance programs examined were less cost–efficient, as only 30 to 60 cents per dollar were spent on successful completions.

SUMMARY FINDING 6:

The vast majority of correctional programs evaluated were cost–effective as, on average, every dollar spent on programming resulted in a return ranging from 1 to 8 dollars. Returns represent cost savings of fewer days of incarceration. This reduction in incarceration costs was, in turn, attributed to earlier releases and extended stays in the community.

SUMMARY FINDING 7:

Correctional programs continue to address the risk levels and specific criminogenic needs of the current offender population, and thus continue to be a relevant, effective means of enhancing public safety and reducing re–offending. Nonetheless, CSC is facing challenges regarding the effective and efficient delivery of correctional programs which will address the increased proportion of offenders with education and learning deficits, mental disorders, shorter sentences, and higher security level ratings.

 

Recommendations

RECOMMENDATION 1:

Correctional Service Canada should further support the implementation of correctional programs by ensuring that: (a) information required for the program referral process is obtained from police and courts at an offenders admission, and entered into the Offender Management System in a timely and efficient manner; (b) the intake assessment process is supported by dedicated staff members; (c) offender program assignments are based on sound referral criteria; (d) referrals are prioritized to target multiple risk factors; (e) programs are effectively offered within a time period limited by sentence length and in a restricted maximum security institutional setting; and (f) offenders are encouraged and motivated to engage in their correctional plans.

RECOMMENDATION 2:

To maximize outcomes attributed to correctional programming, CSC should prioritize programs based on level of risk, offer programs to offenders at intensity levels that are commensurate to sound referral criteria ratings, and develop a strategy to address the violence prevention programming needs of those offenders who also have substance abuse needs.

RECOMMENDATION 3:

Given the positive results of CSC programs in terms of effectiveness, cost–efficiency, and cost–effectiveness, CSC should maximize its return on program expenditures by allocating resources to facilitate the early identification of programming needs, refer more offenders to programs commensurate with their prioritized programming needs, and increase participation rates of offenders who are referred to a program yet do not enroll. Considering the positive outcomes for all offenders participating in community maintenance programs, CSC should also consider strengthening this strategy such that there are more successful completions.

RECOMMENDATION 4:

CSC should develop a strategy that addresses the programming needs of the increased proportion of offenders with education and learning deficits, mental disorders, shorter sentences, and higher security level ratings.

Statistical Overview: Highlights of Selected Findings

The following statistical overview is presented to highlight the many findings contained in this evaluation report. It is important to note that the statistics presented do not represent the sum of all findings contained in the report. All statistics may be found in Appendix J of this report.

The evaluation report examined violent offender programs, family violence programs, substance abuse programs, sex offender programs, living skills programs and community maintenance programs. Analyses focused on specific evaluation objectives: implementation, effectiveness, cost–efficiency, cost effectiveness, and continued relevancy of correctional programs. Below is an overview of results under each of these evaluation objectives; where feasible, results are disaggregated for non–Aboriginal male offenders, Aboriginal male offenders and women offenders.

Implementation

The proportion of offenders who were referred to a program and did not participate varied across program, and by demographic of the federal offender population. Most offenders who were referred to a program and did not participate had short sentences and were serving their sentence in a maximum security institution. Staff members attributed non–enrolments to offenders refusing to participate, a lack of time to complete programs, and a lack of program availability among other reasons.

 


Staff Members Responses: Reasons why an offender might not participate in a correctional program that is included in his or her correctional plan
Reasons why an offender might not participate in a correctional program
Click to enlarge

[D]

Analysis of who enrolled and who did not enrol in Violent Offender Program Referrals.  Statistics for Non-Aboriginal Men, Aboriginal Men, and Women.
(Click to enlarge)

 

Analysis of who enrolled and who did not enrol in Family Violence Prevention Program (High Intensity) Referrals.
(Click to enlarge)

[D]

Analysis of who enrolled and who did not enrol in Substance Abuse Program (High Intensity) Referrals.
(Click to enlarge)

[D]

Analysis of who enrolled and who did not enrol in Sex Offender Program Referrals.
(Click to enlarge)

[D]

Effectiveness, Efficiency and Cost Effectiveness

Positive institutional and community correctional outcomes were revealed for many of CSC’s correctional programs. Program participants were compared to their respective comparison groups12 while controlling for time, risk, and opportunity, across a series of institutional and community outcomes. A ‘positive’ correctional outcome was observed when correctional program participants were found to be i) more likey than than their comparison group to receive a discretionary release, and ii) less likey to return to provincial or federal custody, either with or without (technical revocation) a new offence. This comparsion is illustrated in the first of the three columns in the series of the charts below.

Correctional programs were delivered cost–efficiently. The majority of each dollar spent on correctional programs was spent on program completions, as shown in the second column for each series below. Finally, many correctional programs were assessed as being cost–effective, as positive correctional outcomes resulted in cost savings in terms of avoided incarceration costs.

Male Offenders Violence Prevention Program (VPP)

Positive institutional and community correctional outcomes were revealed for many of CSCs correctional programs.
(Click to enlarge)

[D]

In Search Of Your Warrior (ISOYW)

Proformance and Cost Outcomes for ISOYW
(Click to enlarge)

[D]

High–Intensity
Family Violence Prevention Program (HI–FVPP)

Effectivness Outcomes (either less likely or more likely) for HI-FVPP
(Click to enlarge)

[D]

Moderate–Intensity Family Violence Prevention Program (MI–FVPP)

Effectivness Outcomes (either less likely or more likely) for MI-FVPP
(Click to enlarge)

[D]

High–Intensity Substance Abuse Program (NSAP–High)

Effectivness Outcomes (either less likely or more likely) for NSAP-High
(Click to enlarge)

[D]

Moderate–Intensity Substance Abuse Program (NSAP–Moderate)

Effectivness Outcomes (either less likely or more likely)
(Click to enlarge)

[D]

Aboriginal Offender Substance Abuse Program (AOSAP)

Effectivness Outcomes (either less likely or more likely)
(Click to enlarge)

[D]

High–Intensity Sex Offender Program (HI–SOP) Non–Child, All Regions

Effectivness Outcomes (either less likely or more likely) for HI-SOP
(Click to enlarge)

[D]

1 The return on expenditures reflects overall return on expenditures for all sex offender programs

Moderate–Intensity Sex Offender Program (MI–SOP) Non–Child, All Regions

Effectivness Outcomes (either less likely or more likely) for MI-SOP
(Click to enlarge)

[D]

Continued Relevancy

CSC will be facing challenges regarding the effective and efficient delivery of correctional programs which will address the increasing proportion of offenders with education and learning deficits, mental disorders, shorter sentences and higher security level ratings. The impacts of correctional programs on these demographics of the overall male offender population are illustrated below.

Male Offenders – Violent Offender Programs

Community Outcomes for Violent Offender Programs for the Male Offender population and how effective and efficient the program is.
(Click to enlarge)

[D]

Family Violence Prevention Programs

Community Outcomes for Family Violence Prevention Programs for the Male Offender population and how effective and efficient the program is.
(Click to enlarge)

[D]

National Substance Abuse Programs

Community Outcomes for National Substance Abuse Programs for the Male Offender population and how effective and efficient the program is.
(Click to enlarge)

[D]

Sex Offender Programs
Offenders with non–child victims

Community Outcomes for Sex Offender Programs (with non-child victims) for the Male Offender population and how effective and efficient the program is.
(Click to enlarge)

[D]

Sex Offender Programs
Offenders with child victims

Community Outcomes for Sex Offender Programs (with child victims) for the Male Offender population and how effective and efficient the program is.
(Click to enlarge)

[D]

Living Skills Programs

Community Outcomes for Living Skills Programs for the Male Offender population and how effective and efficient the program is.
(Click to enlarge)

[D]

Staff members largely perceive existing correctional programs as being responsive to the current offender profile, with the exception of offenders with mental health diagnoses and special needs such as cognitive impairments and low educational attainment.

Staff members largely perceive existing correctional programs as being responsive to the current offender profile
(Click to enlarge)

[D]

Almost all staff felt that more could be incorporated into Correctional Programming that would address the needs of offenders with mental health needs and special needs. Recommendations for improving correctional program outcomes for Aboriginal and Women offenders were also offered by many staff members.

 


Staff members responses: Summary of recommendations for improving correctional program outcomes for offenders with mental health needs


Almost all staff felt that more could be incorporated into Correctional Programing
(Click to enlarge)

[D]

Staff members responses: Summary of recommendations for improving correctional program outcomes for offenders with special needs (e.g., low literacy, learning disabilities)

This chart details the results of the staff members responses
(Click to enlarge)

[D]


Staff members responses: Summary of recommendations for improving correctional program outcomes for Aboriginal offenders


Staff members’ responses:  Summary of recommendations for improving correctional program outcomes for Aboriginal offenders.
(Click to enlarge)

 


Staff members responses: Summary of recommendations for improving correctional program outcomes for Women offenders


Staff members’ responses:  Summary of recommendations for improving correctional program outcomes for Women offenders.
(Click to enlarge)

 

Table of Contents

STATISTICAL OVERVIEW: HIGHLIGHTS OF SELECTED FINDINGS

TABLE OF CONTENTS

LIST OF TABLES

INTRODUCTION

1.0 PROGRAM PROFILE

1.1 Background
1.2 Policy and Legislation
1.3 Correctional Service Canadas Correctional Programs
1.3.1 Violent Offender Programs
1.3.2 Family Violence Prevention Programs
1.3.3 Substance Abuse Programs
1.3.4 Sex Offender Programs
1.3.5 Maintenance Programs
1.4 The Correctional Intervention Process
1.4.1 Intake Assessment and Correctional Planning
1.4.2 Intervention
1.4.3 Decisions
1.5 Governance Structure
1.5.1 Stakeholders
1.6 Planned Results
1.6.1 Accreditation of Correctional Programs
1.7 Profile of Correctional Program Participants
1.7.1 Gender and Ethnicity
1.7.2 Sentence Length and Security
1.7.3 Programming Needs
1.8 Financial Expenditures
1.9 Logic Model
1.9.1 Activities and Outputs
1.9.2 Outcomes

2.0 EVALUATION METHOD

2.1 Scope of the Evaluation
2.2 Sample Composition
2.2.1 Study Groups for Quantitative Analyses
2.3 Data Sources
2.3.1 Automated Data
2.3.2 Key Sources for Qualitative Data
2.3.3 Other Data Sources
2.4 Measures
2.4.1 Supplementary Intake Assessment Instruments
2.4.2 Programming Needs Assessment Measure
2.4.3 Measures of Pre–/Post–Program Belief and Attitude Changes
2.4.4 Cost of Maintaining an Offender (COMO)
2.4.5 Cost Efficiency
2.4.6 Cost–Effectiveness
2.4.7 Semi–Structured Interviews
2.5 Institutional and Community Outcomes
2.5.1 Pre–/Post–Correctional Program Treatment Change
2.5.2 Violent Institutional Incidents
2.5.3 Non–violent Institutional Incidents
2.5.4 Substance–related Institutional Incidents
2.5.5 Discretionary Release
2.5.6 Any Readmission
2.5.7 Readmission with a New (Federal or Provincial) Offence
2.5.8 Readmission with a New (Federal or Provincial) Violent Offence
2.5.9 Readmission with a New (Federal or Provincial) Sex Offence
2.6 Procedures
2.7 Analyses

3.0 KEY FINDINGS

3.1 Objective 1: Relevancy:
3.2 Objective 2: Implementation:
3.2.1 Referral Criteria
3.2.2 Supplementary Intake Assessment Tools
3.2.3 Staff Recommendations
3.2.4 Program Referrals and Enrolments
3.2.5 Violent Offender Program Referrals and Enrolments
3.2.6 Family Violence Prevention Program Referrals and Enrolments
3.2.7 Substance Abuse Program Referrals and Enrolments
3.2.8 Sex Offender Program Referrals and Enrolments
3.2.9 Collection and Entering of Information Into OMS
3.2.10 Level of Information Provided to Staff
3.3 Objective 3: Effectiveness
3.3.1 Violent Offender Program
3.3.2 Family Violence Prevention Programs
3.3.3 National Substance Abuse Programs: Overlap of Programming Needs
3.3.4 National Substance Abuse Programs: Effectiveness
3.3.5 Sex Offender Programs
3.3.6 Living Skills Programs
3.3.7 Community Programs
3.4 Objective 4: Cost Efficiency:
3.5 Objective 5: Cost–Effectiveness:
3.6 Objective 6: Continued Relevancy:
3.7 Objective 7: Unintended Findings:
Appendix A: Detailed Program Descriptions
Appendix B: Evaluation Matrix
Appendix C: Referral Criteria for Correctional Programs and Corresponding Variables in Used in the Approximated Measure
Appendix D: Intensity, Duration and Setting of Correctional Programs
Appendix E: Generic Program Performance Measure (GPPM)
Appendix F: Interview Keys
Appendix G: Recalibration and Validation of the Spousal Assault Risk Assessment (SARA) as an Actuarial Tool
Appendix H: Pre–/Post–Program Assessment Battery Results
Appendix I: Staff Responses & Recommendations
Appendix J: Summary of Data
Appendix K: Regional Distribution on Those Referred and Enrolled in Correctional Programs
Appendix L: Schedule I Offences
Appendix M: References

List of Tables

Table 1: Total Expenditures of Correctional Programs

Table 2: Number of Staff Respondents in Each Region, by Location

Table 3: CSC Contributions to Government of Canada Outcome

Table 4: Community Correctional Outcomes for Violent Offender Program Non–Participants

Table 5: Area Under the Curve for the SIR–R1 and the Overall Static Risk Factor Level Rating

Table 6: Community Correctional Outcomes for Family Violence Prevention Program Non–Participants

Table 7: Area Under the Curve for the Spousal Assault Risk Assessment

Table 8: Community Correctional Outcomes for Those Meeting Substance Abuse Program Referral Criteria as per the CASA: Non–Participants

Table 9: Areas Under the Curve for Community Correctional Outcomes: CASA versus the Overall Substance Abuse Need Level Rating: Non–Participants

Table 10: Community Correctional Outcomes for Those Who Met the Approximated Sex Offender Program Referral Criteria: Non–Participants

Table 11: Area Under the Curve for the Sex Offender Priority Rating

Table 12: Ratings of Staff Agreement Regarding the Use of Supplementary Assessment Tools in Decision–making

Table 13: Staff Respondents Who Said Yes to Various Indicators of Programs Training

Table 14: Reasons Listed for Delivering a Program in Which Staff Members Were Not Trained to Deliver

Table 15: Staff Recommendations to Improve Efficiency of Pre–/Post–Assessment Process

Table 16: Current NSAP Referral Criteria (CSC, Reintegration Programs Branch, 2008)

Table 17: Most Common Recommendations to Improve Efficiency of OIA Process

Table 18: Recommendations for Improving the OIA Most Frequently Suggested by Staff Members

Table 19: Effectiveness of Violence Prevention Program for Non–Aboriginal Male Offenders

Table 20: Effectiveness of Violence Prevention Program for Aboriginal Male Offenders

Table 21: Effectiveness of Aggressive Behaviour Control Program for Non–Aboriginal Male Offenders

Table 22: Effectiveness of Aggressive Behaviour Control Program for Aboriginal Male Offenders

Table 23: Effectiveness of In Search of Your Warrior for Aboriginal Male Offenders

Table 24: Effectiveness of Spirit of A Warrior for Aboriginal Women Offenders

Table 25: Effectiveness of High–Intensity Family Violence Prevention Program for non–Aboriginal Male Offenders

Table 26: Effectiveness of Moderate–Intensity Family Violence Prevention Program for non–Aboriginal Male Offenders

Table 27: Programming Needs and Enrolments – Overlap with High Intensity Substance Abuse Programming

Table 28: Effectiveness of NSAP–High for Non–Aboriginal Male Offenders

Table 29: Effectiveness of NSAP–High for Aboriginal Male Offenders

Table 30: Effectiveness of NSAP–Moderate for Non–Aboriginal Male Offenders

Table 31: Effectiveness of NSAP–Moderate for Aboriginal Male Offenders

Table 32: Effectiveness of AOSAP for Aboriginal Male Offenders

Table 33: Effectiveness of NSAP Maintenance Program for Non–Aboriginal Male Offenders

Table 34: Effectiveness of NSAP Maintenance Program for Aboriginal Male Offenders

Table 35: Effectiveness of High–Intensity Sex Offender Programs for non–Aboriginal Male Offenders: Non–Child Victims

Table 36: Effectiveness of High–Intensity Sex Offender Programs for non–Aboriginal Male Offenders: Child Victims

Table 37: Effectiveness of High–Intensity Sex Offender Programs for Aboriginal Male Offenders: Non–Child Victims

Table 38: Effectiveness of High–Intensity Sex Offender Programs for Aboriginal Male Offenders: Child Victims

Table 39: Effectiveness of Moderate–Intensity Sex Offender Programs for non–Aboriginal Male Offenders: Non–Child Victims

Table 40: Effectiveness of Moderate–Intensity Sex Offender Programs for non–Aboriginal Male Offenders: Child Victims

Table 41: Effectiveness of Moderate–Intensity Sex Offender Programs for Aboriginal Male Offenders: Non–Child Victims

Table 42: Effectiveness of Moderate–Intensity Sex Offender Programs for Aboriginal Male Offenders: Child Victims

Table 43: Effectiveness of Low–Intensity Sex Offender Programs for non–Aboriginal Male Offenders: Non–Child Victims

Table 44: Effectiveness of Low–Intensity Sex Offender Programs for non–Aboriginal Male Offenders: Child Victims

Table 45: Effectiveness of Low–Intensity Sex Offender Programs for Aboriginal Male Offenders: Non–Child Victims

Table 46: Effectiveness of Low–Intensity Sex Offender Programs for Aboriginal Male Offenders: Child Victims

Table 47: Effectiveness of Anger & Emotion Management Program for Non–Aboriginal Male Offenders

Table 48: Effectiveness of Anger and Emotion Management Program for Aboriginal Male Offenders

Table 49: Effectiveness of Anger and Emotion Management Program for Women Offenders

Table 50: Effectiveness of Reasoning and Rehabilitation Program for Non–Aboriginal Male Offenders

Table 51: Effectiveness of Reasoning and Rehabilitation Program for Aboriginal Male Offenders

Table 52: Effectiveness of Reasoning and Rehabilitation Program for Women Offenders

Table 53: Effectiveness of Community Maintenance Program for Non–Aboriginal Male Offenders

Table 54: Effectiveness of Community Maintenance Program for Aboriginal Male Offenders

Table 55: Effectiveness of Women Offender Relapse Prevention Program for Women Offenders

Table 56: Cost Effectiveness by Violent Offender Program

Table 57: Cost Effectiveness by Family Violence Prevention Program

Table 58: Cost Effectiveness by Substance Abuse Program

Table 59: Cost Effectiveness by Sex Offender Program

Table 60: Cost Effectiveness by Living Skills Program

Table 61: Summary of Suggestions for Best Practices in the Implementation of Correctional Programs

Table 62: Community Outcomes for All Programs: Male Offenders

Table 63: Community Outcomes for a Sub–sample of Offenders with Less Than Grade 8 Education: Male Offenders

Table 64: Community Outcomes for a Sub–sample of Offenders Identified as Having Learning Disabilities: Male Offenders

Table 65: Community Outcomes for a Sub–Sample of Offenders Diagnosed with Having a Mental Disorder: Male Offenders

Table 66: Ratings of staff agreement as to whether existing correctional programs are responsive to various offender subpopulations

Table 67: Summary of Recommendations for Improving Correctional Program Outcomes for Offenders with Mental Health Needs

Table 68: Summary of Recommendations for Improving Correctional Program Outcomes for Offenders with Special Needs

Table 69: Summary of Recommendations for Improving Correctional Program Outcomes for Aboriginal Offenders

Table 70: Summary of Recommendations for Improving Correctional Program Outcomes for Women Offenders

List of Figures

Figure 1: The Correctional Intervention Process

Figure 2: Stakeholders

Figure 3: Logic Model

Figure 4: Proportion of Offenders Referred to Violent Offender Programs and Not Enrolled, Non–Aboriginal Male

Figure 5: Proportion of Offenders Referred to Violent Offender Programs and Not Enrolled, Aboriginal Male

Figure 6: Proportion of Offenders Referred to SOAW and Not Enrolled, Women

Figure 7: Proportion of Offenders Referred to FVPPs and Not Enrolled, Non–Aboriginal Male

Figure 8: Proportion of Offenders Referred to FVPPs and Not Enrolled, Aboriginal Male

Figure 9: Proportion of Offenders Referred to Substance Abuse Programs and Not Enrolled, Non–Aboriginal Male

Figure 10: Proportion of Offenders Referred to Substance Abuse Programs and Not Enrolled, Aboriginal Male

Figure 11: Proportion of Offenders Referred to SAPs and Not Enrolled, Women

Figure 12: Proportion of Offenders Referred to SOPs and Not Enrolled, Non–Aboriginal Male

Figure 13: Proportion of Offenders Referred to Sex Offender Programs and Not Enrolled, Aboriginal Male

Figure 14: Programming Needs and Enrolments Overlap with High Intensity Substance Abuse Programming: Non–Aboriginal Male Offenders

Figure 15: Programming Needs and Enrolments – Overlap with High Intensity Substance Abuse Programming: Aboriginal Male Offenders

Figure 16: Programming Needs and Enrolments –– Overlap with High Intensity Substance Abuse Programming: Women Offenders

Figure 17: Cost Allocation for Violent Offender Programs

Figure 18: Cost Allocation for Family Violence Programs

Figure 19: Cost Allocation for Living Skills Programs

Figure 20: Cost Allocation for Sex Offender Programs

Figure 21: Cost Allocation for Substance Abuse Programs

Figure 22: Cost Allocation for Community Programs

Figure 23: Cost Allocation for Women Offender Programs

List of Acronyms

A & EM
Anger and Emotions Management Program
ABC
Aggressive Behaviour Control Program
ACCOP
Commissioner’s Directive
ACPA
Correctional Service Canada
ADS
Alcohol Dependence Scale
ALO
Aboriginal Liaison Officer
AOSAP
Aboriginal Offender Substance Abuse Program
AUC
Area Under the Curve
AWI
Assistant Warden Intervention
BHP
Basic Healing Program
CASA
Computerized Assessment of Substance Abuse
CCRA
Corrections and Conditional Release Act
CD
Commissioner’s Directive
CEA
Cost–Effectiveness Analysis
CIB
Correctional Intervention Board
CLAI
Computerized Lifestyle Assessment Inventory
CMLC
Correctional Management Learning Centre
CMP
Community Maintenance Program
COMO
Cost of Maintaining an Offender
CPIC
Canadian Police Information Centre
CPO
Correctional Program Officer
CPPR
Correctional Plan Progress Report
CSC
Correctional Service Canada
DAST
Drug Abuse Screening Test
DOJ
Department of Justice
EMS
Expenditure Management System
EXCOM
Executive Committee
FPS
Finger Print System
FTE
Full Time Equivalent
FVPP
Family Violence Prevention Program
FVRA
Family Violence Risk Assessment
GPPM
Generic Program Performance Measure
HRMS
Human Resource Management Sector
IFMMS
Integrated Financial and Materials Management System
IMRS
Integrated Management Reporting System
ISOYW
In Search of Your Warrior
ISU
Intensive Support Units
ITT
Intent–to–Treat
MBIS
Motivation Based Intervention Strategy
NHQ
National Headquarters
NPB
National Parole Board
NPM
National Program Managersent
NSAP
National Substance Abuse Program
OAG
Office of the Auditor General Canada
OIA
Offender Intake Assessment
OMS
Offender Management System
PAA
Program Activity Architecture
PIDS
Perimetric Intrusion Detection System
PRD
Problems Related to Drinking
RACP
Regional Administrators of Correctional Programs
R & R
Reasoning and Rehabilitation Program
RHQ
Regional Headquarters
ROC
Receiver Operating Characteristic
SAP
Substance Abuse Program
SARA
Spousal Assault Risk Assessment
SAS
Statistical Analysis System
SIR
Statistical Information on Recidivism
SOP
Sex Offender Program
SOAW
Spirit of A Warrior
SSOA
Specialized Sex Offender Assessment
TBS
Treasury Board of Canada Secretariat
VOP
Violent Offender Program
VPP
Violence Prevention Program
WED
Warrant Expiry Date
WOSAP
Women Offender Substance Abuse Program

Introduction

1.0 Program Profile

1.1 Background

Correctional Service Canada13(CSC) delivers structured interventions that address factors directly linked to offenders’ criminal behaviour (Correctional Service Canada [CSC], 2003). The primary goal of these correctional programs is to help offenders safely and successfully reintegrate into the community (Corrections and Conditional Release Act [CCRA], 1992) thereby enhancing public safety. The programs cover diverse content areas and include violent offender programs, family violence programs, substance abuse programs, sex offender programs, and life skills interventions14. Maintenance programs are also offered for several of these areas in order to ensure continuity of care to maximize treatment outcome. In addition, programs addressing the unique and diverse needs characteristic of Aboriginal and women offenders have been developed and implemented, with the aim of improving outcomes for these groups across several program areas.

1.2 Policy and Legislation

CSC is mandated by law to provide these types of programs (CCRA, 1992, s. 5(b)). The CCRA (1992), section 3, paragraph 5(b) and sections 76, 77, 79 and 80 provide the legislative framework guiding the development, implementation, and maintenance of reintegration programming for the Correctional Service of Canada:

3. The purpose of the federal correctional system is to contribute to the maintenance of a just, peaceful and safe society by . . . (b) assisting the rehabilitation of offenders and their reintegration into the community as law–abiding citizens through the provision of programs in penitentiaries and in the community.

In addition to the CCRA, Commissioners Directive [CD] 726: Correctional Programs (CSC, 2003) provides a policy framework from which CSC can implement reintegration programming. The policys objective is two–fold first, to ensure that correctional programs meet the identified needs of offenders and promote successful reintegration, and second, to ensure the integrity and effectiveness of the correctional programs offered to offenders. Specifically, it is stated that offenders are to be assigned to a correctional program based on their correctional plan and on established correctional program selection criteria (CSC, 2003, para. 11). It is also stipulated that offender participation in correctional programs must be voluntary and based on informed consent (CSC, 2003, para. 12).

The development and implementation of a correctional plan is also mandated by section 102 of the CCRA:

102. The institutional head shall ensure that a correctional plan for an inmate is developed as soon as practicable after the reception of the inmate in the penitentiary, and is maintained, with the inmate to ensure that the inmate receives the most effective programs at the appropriate time in the inmate’s sentence to prepare the inmate for reintegration into the community, on release, as a law–abiding citizen.

The focus and delivery of correctional programs is guided by CD 726, paragraph 13 (CSC, 2003). Specifically, correctional programs must: (a) be based on theory and supporting research (an empirically–based model of change); (b) target criminogenic factors; (c) be skills–oriented; (d) take into account the particular characteristics of offenders to help ensure that they derive maximum benefit from the program (responsivity); (e) address the particular risk and need profiles of offenders through their scope, intensity, duration and type of group setting (program intensity); (f) ensure a continuum of care between institutions and the community (continuity of care); (g) include a detailed program description; (h) include a plan for monitoring and evaluation; and (i) be delivered using proven treatment methods, in the least restrictive environment possible consistent with staff, offender and public safety, and according to approved standards.

Programs must also address the diverse needs of women, Aboriginal and other groups of offenders with special requirements (CSC, 2003, para. 13). This is congruent with CCRA section 77 (1992), stipulating that CSC must (a) provide programs designed particularly to address the needs of women offenders, and (b) consult regularly about programs for women offenders with appropriate womens groups, and other appropriate persons and groups with expertise on, and experience in working with, women offenders. Similarly, section 80 of the CCRA (1992) states that CSC must provide programs designed particularly to address the needs of Aboriginal offenders.

1.3 Correctional Service Canadas Correctional Programs

The primary goal of correctional programs is to help offenders safely and successfully reintegrate into the community. These programs involve interventions that focus on the many factors that directly contribute to criminal behaviour, and there is an array of programs to address the wide variety of offender needs. Correctional programs include clearly defined objectives, participant selection criteria, a process for evaluating participant progress, and a process for evaluating the effectiveness of the program. The programs must be delivered by trained and qualified Correctional Program Officers (CPO) or psychologists, and high intensity programs are typically co–facilitated by a CPO and a psychologist. Program descriptions are presented below in the same categories as they are described in CSC’s Program Activity Architecture (PAA).

Women offenders programs are based on research that is sensitive to the unique situation of women offenders. Programs for women must use an approach that responds to the multi–faceted needs of women offenders and address emotional regulation issues which underlie other needs such as cognitive functioning and/or substance abuse.

Aboriginal offender programs are developed and delivered by Aboriginal sources, including Elders, and include cultural teachings and traditions. Aboriginal correctional programs not only teach skills, they impart a pro–social value system to participants that stresses the importance of family, community and history.

1.3.1 Violent Offender Programs

The goal of violent offender programs (VOP) is to improve the skills of participants, and subsequently, to reduce the risk of future violence. The programs are mainly skills–based with an emphasis on relapse prevention. Intervention techniques emphasise self–control, social problem solving, education and self–management. There also exist VOPs designed specifically to address the static and dynamic risk factors of women and Aboriginal men15.

1.3.2 Family Violence Prevention Programs

CSC’s national family violence prevention programs (FVPP) are primarily focused on male offenders who have been abusive in their intimate relationships with female partners or ex–partners. These programs are based on the principle that responding with violence and abuse in intimate relationships is a learned pattern of behaviour that can be modified.  The programs teach participants to understand the dynamics of their abusive relationships and train them in techniques that will allow them to replace their abusive behaviours with approaches that are conducive to forming healthy non–abusive relationships.  As such, treatment targets include problematic thinking, emotional control, social skills, and self–control related to spousal violence and abuse. FVPPs exist for different intensity levels (moderate and high) and for Aboriginal offenders (Aboriginal High Intensity Family Violence Prevention Program (AHIFVPP))16.

1.3.3 Substance Abuse Programs

Substance abuse programs (SAP) endeavour to alter behaviours through reinforcement, modeling and skills acquisition through role play. Participants learn various skills including problem solving, emotional control, social skills and intrapersonal skills or self–management. Substance abuse interventions are offered at varying intensity levels (low, moderate, and high), and for specific sub–populations (women and Aboriginal offenders), and also include institutional and community maintenance programs17.

1.3.4 Sex Offender Programs

Sex offender programs (SOP) are therapeutic cognitive–behavioural interventions based on criminogenic needs and known risk factors associated with individual offenders’ patterns of sexual offending behaviour. The programs emphasize the need for offenders to take responsibility for their actions, recognize the behavioural progression that preceded and followed sex offences, identify situations which place them at risk to re–offend, and assist them to develop strategies to prevent recidivism. In addition to self–management, the program components include cognitive distortions, deviant arousal and fantasy, social skills, anger and emotion management, empathy and victim awareness. Programs are designed for sex offenders convicted of a sexual crime, offenders who were convicted of a sexually motivated crime, or those who have admitted being guilty of a crime of a sexual nature for which they have not been convicted. Two programs are delivered nationally: the Moderate Intensity Sex Offender Program (MI–SOP) and the Low Intensity Sex Offender Program (LI–SOP). Regionally developed high intensity programs exist in the Quebec, Ontario, Prairie, and Pacific Regions. Finally, CSC also offers a SOP for women offenders18.

1.3.5 Living Skills Programs

Reasoning and Rehabilitation Program

The Reasoning and Rehabilitation (R & R) program is an accredited moderate intensity program that teaches male and female offenders skills that address the socio–cognitive deficits linked to criminal behaviour. Specific target areas include self–regulation and self–management, egocentrism/lack of social perspective, problems related to assertiveness and problematic social interactions, poor critical reasoning, and rigid thinking. The program consists of 37 group sessions each lasting 2 to 3 hours and four individual sessions. It is offered in institutions and in the community. Following the program, male and female offenders who still have significant socio–cognitive deficits are directed to the Cognitive Skills Maintenance Program.

Anger and Emotions Management Program

The Anger and Emotions Management Program (A & EM) is an accredited moderate intensity program for offenders who are assessed at moderate or high risk to re–offend, or who have low to moderate reintegration potential.  Offenders’ criminal behaviour must be associated with inappropriate management of their emotions; offenders must also show some or considerable need in the personal/emotional domain, and score positively on at least one of the following indicators: (a) aggression; (b) assertion problem; (c) poor conflict resolution; (d) hostility problem; (e) poor stress management; (f) low frustration tolerance; (g) worries unreasonably; (h) problematic risk taking; and (i) thrill seeking.   It is offered to male and, by using an adapted version, women offenders. The program teaches offenders to change their thinking patterns that trigger and feed emotions associated with their criminal behaviour.  The program consists of 27 group sessions, each lasting 2 to 3 hours, and also includes two individual sessions.  It is available in institutions and in the community. Following the program, offenders who still have problems are directed to the Anger and Emotions Management Maintenance Program.

1.3.5 Maintenance Programs

As a response to logistic problems encountered by CSC, such as the need for group settings, locations of offenders in the community and access to groups, and the wide variety of maintenance programming needs, a generic Community Maintenance Program was developed in an attempt to meet the needs of all offenders. The generic maintenance program focuses on common group themes, such as following a relapse prevention plan and avoiding high risk situations. Prior to this generic model, there was an array of program–specific maintenance programs available to offenders, offered in both the institution and the community. Each maintenance program was designed to reduce offenders risk by maintaining skills and benefits learned while in specific correctional programs19.

1.4 The Correctional Intervention Process

The correctional intervention process is organized into three main components (see Figure 1):

  • Intake assessment and correctional planning
  • Intervention with the offender
  • Decision process

1.4.1 Intake Assessment and Correctional Planning

Information gathering starts as soon as the offender has been sentenced. The Preliminary Assessment is used to collect basic data on the offender, assess his or her immediate needs, and initiate the collection of critical documents. These critical documents (police and court records, psychological assessments etc.) are used to populate the supplementary intake assessments which, in turn, yield programming needs assessments. Ultimately, the programming needs assessments guide offender referrals to appropriate correctional programs. The programming needs assessment focuses mainly on dynamic contributing factors; it identifies the various interventions required (programs, intensity, other intervention activities), the sequence, the location (institution, security level, community), and the appropriate timing as a function of critical dates (eligibility dates). It also stipulates the outward indicators of change, which allows for a conclusion that change has occurred.

1.4.2 Intervention

Figure 1: The Correctional Intervention Process The Correctional Plan is the road map for intervention.

[D]

The Correctional Plan is the road map for intervention. Through this process, programming needs assessments facilitate program referrals and, subsequently, program enrolments. In institutions, the purpose of correctional programs is to mitigate those dynamic factors contributing to criminal behaviour, through the strict implementation of the Correctional Plan. This implies a regular review of the offender’s progress through analysis of observations, the casework record, program report, specialists assessments and discipline file in case conferences. As a result of this progress review, intervention activities can be adjusted in light of the results recorded (aftercare, interviews, new programs, etc). The results of this review are recorded in the Correctional Plan Progress Report (CPPR).

The CPPR allows for following the evolution of the offender, including re–evaluation of the level of intervention required, the reintegration potential, the relevance of the security classification, and marking the progress accomplished by the offender. To evaluate progress, a set of indicators of change must be examined. Change is the acquisition of new skills, knowledge, abilities and insight into behaviour. Once offenders have gained new skills or knowledge, they must integrate them into their behaviour. It is not always necessary to attain this level of change within the institution. If the risk is assumable, this integration can be done in the community. This phase is very important for the protection of society in the long term.

Following participation in programming, offenders should be referred to maintenance programming, if required. Any exception to this guideline should be documented in the Correctional Plan or the CPPR after consultation with a program specialist in the program area. Maintenance programs are offered in the institution and in the community. Once an offender has completed a program, Institutional and Community Parole Officers can make a referral to a maintenance program as needed in consultation with a program specialist in the program area. Level of participation in maintenance programming is based on risk, criminogenic needs, and ability to manage risk to re–offend. An offender may repeat a maintenance program a number of times in accordance with the recommendations of the case management team and the program facilitator.

1.4.3 Decisions

The CPPR is an integral part of the documentation submitted for decision–making. All information used in developing the offender’s profile, the Correctional Plan, and the CPPR is placed on file. Decisions constitute critical points in the sentence; this is the point at which the location of placement, transfer and all forms of release are determined.

1.5 Governance Structure

1.5.1 Stakeholders

The Executive Committee (EXCOM), whose members primary responsibility is to provide overall governance, strategic planning, policy, and oversight of policy implementation and general processes within CSC, authorizes the development of new correctional programs based on the recommendation of the Assistant Commissioner, Correctional Operations and Programs (ACCOP; CSC, 2003, para. 6). At National Headquarters (NHQ), the ACCOP is responsible for the management, coordination and monitoring of correctional program development and delivery, the management of the program and site accreditation processes, and the training and certification of Program Trainers and CPOs (CSC, 2003, para. 7). In addition, each Regional Headquarters (RHQ) must ensure that an annual program plan is developed for each operational unit and that the infrastructure for the administration and the delivery of correctional programs meets the needs of offenders. At the institutional level, each Institutional Head/Warden and District Director is responsible for the implementation of the annual program plan, offender enrolment and program delivery, as well as for ensuring that data related to program evaluation are collected and available for research and monitoring purposes. In addition, program awareness sessions are to be provided on a yearly basis and prior to the introduction of a new program to those operational staff who interact with offenders. Finally, all staff members who work directly with offenders are responsible for the ongoing sharing of information regarding the offender’s program participation and progress against criminogenic needs (CSC, 2003, para. 10). For a flowchart of stakeholders in correctional programming, see Figure 2.

Figure 2: Stakeholders The Executive Committee (EXCOM), whose members primary responsibility is to
(Click to enlarge)

[D]

1.6 Planned Results

Based on an evaluation of the above–listed correctional programs, it is expected that: (a) programs address the criminogenic needs of the current offender profile; (b) offenders who are referred to the appropriate intensity level of programming according to their risk and needs show increased positive correctional outcomes in the institution and community; and (c) programs are delivered in an efficient and cost–effective manner. It is important to note that although important, this evaluation does not assess the design of correctional programs, nor the degree to which they are delivered within acceptable standards. Correctional program design and delivery is assessed through the program accreditation process (described in the proceeding section). A detailed list of expected results for each evaluation objective is provided in the Scope of the Evaluation section.

1.6.1 Accreditation of Correctional Programs

It is important to know whether the programs are delivered as intended and the offenders are changing as intended. To ensure that its programs are designed to maximize effectiveness and embrace state–of–the–art treatment techniques and delivery standards, CSC has adopted a set of accreditation standards. Programs are presented to review panels of internationally–recognized experts in the field who assess the design of the programs, and the manner in which they are delivered, in relation to eight criteria.

The first three criteria require the programs to be grounded on theory. Programs that provide an explicit explanation of how they should affect change and that can justify this explanation based on research are most likely to be effective. In addition, effective programs change problems directly linked with offending (i.e., target criminogenic needs) and employ methods of delivery with demonstrated value. The following two criteria assess essential elements of the programs design and content. They specify that the programs should involve teaching offenders skills to help offenders avoid criminal activities and successfully engage in legitimate pursuits and be attentive to their learning style and abilities. Criteria six and seven concern the offenders criminogenic needs and subsequent portions of their sentence. To be effective, programs should be of sufficient intensity and duration to adequately address the range and severity of the problems typical of their participants and include provisions for follow–up care in the community.

The final criterion focuses on issues related to continuous review and stipulate that the programs need to be continuously monitored and evaluated to ensure integrity.

1.7 Profile of Correctional Program Participants

Program participants were profiled to gain an understanding of their demographic information, and treatment needs. The sample represents all offenders warrant of committal admissions between 1997 and 2007 (N = 48,601).

1.7.1 Gender and Ethnicity

The vast majority of offenders were male (95%; N = 46,038 compared to 5%; N = 2,563 women). Of the men, 82% (N = 37,570) were non–Aboriginal and 18% (N = 8,183) were Aboriginal. The proportion of Aboriginal women was slightly higher than that of the men, with 26% (N = 647) of women offenders being Aboriginal and 74% (N = 1,864) being non–Aboriginal. The remaining profile statistics will be presented for Aboriginal males, non–Aboriginal males and women offenders. There was some attrition due to the lack of information regarding ethnicity, where if this variable was missing, offenders were not include in further analysis that grouped Aboriginal and non–Aboriginal offenders.

1.7.2 Sentence Length and Security

Non–Aboriginal Male Offenders

The average sentence length for non–Aboriginal male offenders, excluding those who have received a life sentence is 3.7 years. The highest proportion of offenders receive a medium security rating (62%; n = 12,031), followed by minimum security (30%; n = 5,815) and maximum security (8%; n = 1,567).

Aboriginal Male Offenders

The average sentence length for Aboriginal male offenders, excluding those who have received a life sentence was 3.56 years. The highest proportion of Aboriginal offenders receive a medium security rating (74%; N = 3,147), followed by minimum security (15%; N = 622) and maximum security (11%; N = 477).

Women Offenders

The average sentence length for women offenders, excluding those who have received a life sentence was 3.15 years. The highest proportion of women offenders receive a minimum security (50%; N = 697) rating followed by medium security rating (43%; N = 601), and maximum security (6%; N = 89).

1.7.3 Programming Needs

The referral criteria that CSC uses to determine programming needs were approximated using offender data available from the Offender Intake Assessment (OIA), the Canadian Police Information Centre (CPIC) data and supplementary assessments conducted at offender admission. Details regarding how these criteria were developed and applied are in Appendix C.

Non–Aboriginal Male Offenders

More non–Aboriginal male offenders met the referral criteria to participate in substance abuse than any other program area. Only 34% (N = 11,958) of the entire non–Aboriginal male correctional population studied did not meet the criteria for any substance abuse program. The most frequently met criteria for SAPs was the low intensity (28%; N = 9,723) followed by the moderate intensity program (20%; N = 6,856), and high intensity (19%; N = 6,546). Many non–Aboriginal male offenders also met the criteria for the violent offender programs (VOP). Close to one–quarter (23%) of all offenders in the sample met the criteria for participation in the High Intensity Violence Prevention Program (VPP). For those offenders who meet the criteria for FVPP, more offenders met the criteria for participation in moderate intensity programs (10%; N = 2,063) compared to the high intensity program (6%; N = 1,274).

Aboriginal Male Offenders

More Aboriginal male offenders met the criteria for SAPs than any other program area. Only 14% (N = 1,008) of the sample did not meet the criteria for a SAP. More Aboriginal offenders met the criteria for the High Intensity Substance Abuse Program (NSAP–High; 31%; N = 2,289), followed by moderate intensity (29%; N = 2,117), and low intensity (27%; N = 2,005). Almost half of all Aboriginal offenders also met the criteria for the high intensity VPP (44%; N = 3,590). For FVPPs, more offenders met the criteria to participate in the moderate intensity (18%; N = 802) than the high intensity (14%; N = 595). Lastly, one–quarter of Aboriginal offenders meet the criteria for SOPs.

Women Offenders

A large proportion of women offenders met the referral criteria for a SAP program. Approximately one–third were rated as high intensity (35%; N = 762), while smaller numbers also needed moderate intensity (16%; N = 338) and low intensity (21%; N = 452). Using the violent prevention program criteria currently being used for the recently implemented womens VOP, 7% (N = 186) of the women offender sample also required high intensity violence prevention programming.

1.8 Financial Expenditures

The following table illustrates the total program expenditures for fiscal years 2005/2006 and 2006/2007. Total expenditures include salary dollars to administrator programs and Operational and material management dollars.

Table 1: Total Expenditures of Correctional Programs
   Fiscal Year 2005/2006
Expenditure
 Fiscal Year 2006/2007
Expenditure
 All Programs20  $53,232,479  $52,309,418
Correctional Programs21  $26,345,107  $25,947,480
 Sex Offender Programs22  $4,746,022  $4,391,955
 National Moderate Intensity Sex Offender  $922,580  $618,471
 National High Intensity
National Low Intensity Sex Offender  $170,852  $72,485
 Maintenance Sex Offender  $307,870  $471,061
 Family Violence  $2,330,671  $2,420,430
 High Intensity Family Violence  $553,869  $575,032
 Maintenance  $50,676  $1,816
 Moderate Intensity Family Violence  $506,926  $549,783
 Aboriginal High Intensity Family Violence  $0  $0
 Living Skills Programs  $4,489,750  $4,483,922
 Reasoning and Rehabilitation (discontinued this year)  $2,494,164  $2,182,093
 Cognitive Maintenance  $268,890  $192,700
 Anger and Emotions Management  $1,226,090  $1,403,973
 Anger and Emotions Management Maintenance  $30,639  $20,106
 Circles of Change  $54  $36,537
 Aboriginal Basic Healing Program  $189,784  $358,337
 Community Correctional Programs  $1,359,201  $1,248,248
 Counter Point  $1,184,936  $925,865
 Community Maintenance  $174,265  $322,383
 Violent Offender Programs  $3,866,253  $3,300,786
 Violence Prevention Program  $1,546,764  $1,311,349
 In Search of Your Warrior – High Intensity  $42,834  $116,751
 Spirit of A Warrior  $523,604  $505,848
 Substance Abuse Programs  $6,991,937  $7,499,484
 National Substance Abuse Program – High  $1,717,724  $2,003,417
 National Substance Abuse Program – Moderate  $2,349,381  $2,777,941
 National Substance Abuse Program – Low  $821,997  $591,348
 National Substance Abuse Maintenance  $773,657  $793,306
 Women Offender Substance Abuse  $784,690  $642,698
 Aboriginal Offender Substance Abuse Program  $75,911  $95,270
 Women Offender Programs  $622,771  $546,149
 Women Offender Relapse Prevention  $160,790  $109,757

 

1.9 Logic Model

The logic model is a schematic representation of the essential components of that which is being evaluated (Posovac & Carey, 1992). It delineates the philosophy and relevant issues that underlie correctional programs (see Figure 3).

1.9.1 Activities and Outputs

The activities illustrated in the logic model represent activities in which the correctional program stakeholders engage to achieve the goals of the strategy. Four activities are identified in the logic model, and are essential to meeting the requirements of the program strategy.

Activities involved in the Correctional Program Strategy are along two streams: offender activities and CPO activities. The offender–centered activities begin with the OIA. From the OIA, offenders will receive more specialized risk assessment if needed (e.g., substance abuse or family violence assessment). Based on these assessments, offenders are referred to a program or several programs via the Correctional Intervention Board (CIB). Program referrals also depend on how each site prioritizes offenders’ needs. The resulting product of these activities is the number of program referrals based on offenders’ risk/needs profiles and assessed intensity levels.

The CPO activities include the selection and training of CPOs, and refresher training and/or continued training toward certification. The resulting output from this training is the number of CPOs who are trained to deliver correctional programs.

1.9.2 Outcomes

Outcomes are changes resulting from the activities carried out. As illustrated in the logic model (Figure 2), outcomes are grouped into immediate, intermediate, long–term, and ultimate goals.

Immediate Outcomes

The immediate outcome of program referrals and CPO training is participation in correctional programs that are delivered by trained CPOs. For those offenders who are not ready for participation in programs, participation in pre–program activities (e.g., Motivational Interviewing or the Basic Healing Program) aims to increase their readiness for program participation.

Intermediate Outcomes

Intermediate outcomes will also be measured as part of the evaluation, and are based on the assumption that offenders complete a program, either in the institution or the community. The first of the intermediate outcomes is that offenders targeted dynamic needs are addressed; this includes the acquisition of cognitive and social abilities, an understanding of and the ability to explain the offence cycle, improved management and understanding of emotions, and the ability to explain and use what has been learned. In addition, more positive decisions for discretionary release due to reduced risk will be measured. The last intermediate outcome that will be examined is increased safety in the institution/community for the offender population and CSC staff, as indicated by reduced incidents.

Long Term Outcome

The long-term outcome to be evaluated is the consistent and generalized use of acquired skills and knowledge in interactions, and the ability to perceive and change ones own offence cycle. These changes are indicated by reduced revocations/recidivism, and result in improved reintegration into the community. This is differentiated from intermediate outcomes in that the offender uses his or her skills and knowledge more consistently and generally, and not only in specific settings or situations. This consistent use of skills contributes to an overall improved reintegration into the community.

Ultimate Outcome

The ultimate outcome of the Correctional Program Strategy is that offenders are safely and effectively accommodated and reintegrated into Canadian communities with due regard for public safety.

 

Figure 3: Logic Model The logic model is a schematic representation of the essential components of that which is being evaluated (Posovac & Carey, 1992).
(Click to enlarge)

[D]

2.0 Evaluation Method

2.1 Scope of the Evaluation

The evaluation was conducted by the Evaluation Branch, CSC, and various components were reviewed by CSCs Reintegration Programs Branch and Research Branch. This targeted evaluation was completed to inform decision–making on the future disposition of the programs and focused on success, cost–effectiveness, continued relevancy, as well as unintended effects. Key results, performance indicators, analyses and data sources for each of these evaluation objectives are listed in the Evaluation Matrix (see Appendix B).

At the outset of the current evaluation, the following expected results were identified under each evaluation objective:

Objective #1: Continued Relevancy
  1. Changes to programs have been made in accordance with the changing offender profile.
  2. Programs are positioned to address future challenges given upcoming legislation.
Objective #2: Program Implementation
  1. Offenders with identified needs are being referred to the recommended programs.
  2. Offenders with identified needs are being enrolled in the recommended programs.
  3. Participation in accredited programs is associated with better outcomes relative to participation in non–accredited programs.
Objective #3: Success (Effectiveness & Efficiency)
  • Correctional programs improve beliefs and attitudes toward treatment targets.
  • Program participation leads to a reduction in offenders involvement in major and minor institutional incidents.
  • Those who participate in programs are deemed to be more manageable in the community, and are thus more likely to be granted a discretionary release by the National Parole Board (NPB).
  • Offenders who have taken programs that address their needs are:
    • more likely to stay in the community longer;
    • less likely to return with a technical revocation;
    • less likely to return with a new offence;
    • more likely to reach the end of their sentence; and
    • more likely to remain in the community past the end of their sentence.
  • There are sufficient numbers of CPOs and Aboriginal CPOs to deliver the required menu of correctional programs at each site.
  • Training resources are appropriately directed toward staff, so that trained staff will deliver programs to offenders.
  • Training for CPOs is repetitive to the extent that generic program delivery modules are repeated across multiple program–specific training courses.
  • Gains in efficiency would be made if integrated program content was delivered to offenders prior to participation in program–specific modules
Objective #4: Cost–effectiveness
  1. Correctional program expenditures yield positive institutional and community outcomes that are in turn associated with financial returns.

 

2.2 Sample Composition

2.2.1 Study Groups for Quantitative Analyses

In order to analyze group differences with regard to institutional and community outcomes, CSCs Offender Management System (OMS) data were used to identify offenders program assignment status for a sample of first–term federal admissions between the years 1997 and 2007 (N = 48,601). Subsequently, two groups were created, an Intent–to–Treat (ITT) group of offenders who were exposed to the targeted correctional program through varying degrees of program participation, and a comparison group of offenders who were assigned to a program but in which they did not participate. Descriptions of both groups are listed below:

Intent–to–Treat Group

For analytical purposes, three treatment sub–groups comprised an ITT group; program completers, program non–completers (due to population management issues) and those who dropped out of the program. Details regarding each ITT subgroup are as follows:

 

Completed the program

Successful. Offenders who completed the program and showed acceptable attendance and an indication of treatment gain, as recorded on the final program evaluation.

Attended all sessions. Offenders who completed the program, but for whom there was no evidence of treatment gain. This assignment status usually results in a negative final assessment of the offenders participation in the program, or an indication that the offender failed to demonstrate that he/she has internalized all of the skills of the program.

Unsuccessful. Formerly used to refer to those offenders who attended all sessions, but who did not show treatment gain (see above).

 

Population management

Transferred. Offenders who were transferred outside of their current facility, but who were assigned to and participated in a program immediately prior to their transfer from the facility. Since the program was never completed, this will indicate the status of the offenders involvement. It also applies in the community setting where an offender is transferred to a different supervision office or location and the current program participation ends with the transfer of supervision.

Assignment transferred. Offenders who have been assigned or temporarily reassigned to a program, and who permanently leaves the program to attend another assignment at the same site.

Warrant Expiry Date (WED) reached. Offenders who were participating in a program when they reached their WED.

Released. Offenders who were assigned to and participated in a program up to and immediately prior to the time of a scheduled release on day parole, full parole, or statutory release.

Temporarily reassigned. Offenders who have left their current program assignment in order to participate in a different assignment for a short period of time, with the expectation that they will return to the original assignment.

Program cancelled.Offenders who were participating in a program and the program was cancelled while the offender was enrolled. The cancellation of the program was due to circumstances beyond the control of the offender.

Paroled. Formerly indicated offenders who were assigned to a program, and who were released on day parole or full parole during their participation in that program.

 

Drop–Outs

Incomplete. Offenders who were participating in a program but who did not complete the program or were removed from the program. Typically this status is used when an offender is admitted to segregation or removed from the program for a non–program related offence, or for outside court or hospital reasons. This status is also used for offenders who are under supervision in the community who have not completed their program due to suspension of their conditional release for reasons unrelated to their participation in the program.

Suspended. Offenders who were enrolled in the program and will not be readmitted without a new placement decision by the Program Board. This is normally used to identify disciplinary concerns that have arisen during the program, or when the offenders quality of participation does not meet participation, behavioural, or other program–related standards and expectations. This also includes offenders who quit the program.

Suspension pending. Formerly indicated offenders whose participation in a program was under review by the Program Assignment Board, following the program facilitators decision to suspend the offender from the program. This is a temporary status that would result in either suspension or temporary suspension.

Temporary suspension. Formerly indicated offenders who were suspended from a program with the agreement that the offender would return to the program after an agreed period of time. This status has since been replaced with temporarily reassigned (defined earlier) to indicate that the offender started the program and is actively involved with the program, with the expectation that the offender will return to the program.

Comparison Groups

A pool of offenders was identified in order to facilitate comparisons on institutional and community outcomes, while controlling for time, risk and opportunity. Comparison groups may have varied depending on the target program.

Assignment cancelled

The group used as the primary pool of candidates to draw from when making comparisons with the ITT group were those whose application or referral to a program was processed by the Program Assignment Board (or equivalent), and were put on the wait list for that program. Prior to actually participating in the program, a decision was made to not proceed with the offenders assignment to the program. Statistical analyses incorporated control measures in order to control for static and dynamic risk and opportunity. The data were disaggregated by gender and, within the male population, Aboriginal and non–Aboriginal offenders. As such, when comparisons were made for the non–Aboriginal male, Aboriginal male and women offender program participants, comparisons were always drawn from those respective comparison sub–groups, using statistical procedures which controlled for time, opportunity and risk.

Matched Groups

One–to–one matched grouping was used for samples which were small or in which the majority of those referred to the program enrolled in that program. This technique was used for the analyses for the womens VOP (Spirit Of A Warrior (SOAW) and SAP (Women Offender Substance Abuse Program (WOSAP). Groups were matched on static and dynamic risk factors, including overall risk, overall need in the employment/education, family–marital, associates, substance abuse, community functioning, personal emotional and attitude domains of the OIA23, as well as age at admission, and sentence length.

2.3 Data Sources

2.3.1 Automated Data

Data from CSCs automated database OMS was extracted for warrant of committal admissions to federal custody from January 1st, 1997 to November 15th, 2007 (N = 48,264). The sample was composed of 37,570 (78%) non–Aboriginal male offenders, 8183 (17%) Aboriginal male offenders, and 2,511 (5%) women offenders. It should be noted that, due to the long time period covered by this data, individuals with multiple sentences may be reflected more than once.

The key data source of financial information, used for the cost–effectiveness analyses, was drawn from CSCs Integrated Management Reporting System (IMRS).

2.3.2 Key Sources for Qualitative Data

Information was collected from individuals and groups considered to be primary stakeholders as specified in CD 726 (CSC, 2003). Overall, 242 stakeholders were interviewed.

Stakeholders from institutions across all five regions (Atlantic, Quebec, Ontario, Prairie, and Pacific) were interviewed, in addition to staff from Community Parole Offices and RHQs. Staff reported having been in their current position anywhere from 1 month to 32 years, with an average duration of 5.68 years (SD = 5.97). Representation of interview respondents by location and region is presented in Table 2.

 

Table 2: Number of Staff Respondents in Each Region, by Location
  Region  
Pacific Prairie Ontario Quebec Atlantic Total
Location
   Institution 34 26 30 36 41 167
   Community 8 10 7 10 11 46
   RHQ 4 4 3 6 6 23
   Unspecified 2   1   1 4
Total 48 40 41 52 59 240

Note: Two respondents (0.8%) did not identify either their location or region.

 

Stakeholders interviewed from regional offices (9.7%, N = 23) included Regional Program Managers (N = 17), Regional Administrators of Correctional Programs (N = 3), and Assistant Deputy Commissioners (N = 3).

Stakeholders interviewed from regional offices (9.7%, N = 23) included Regional Program Managers (N = 17), Regional Administrators of Correctional Programs (N = 3), and Assistant Deputy Commissioners (N = 3).

Community respondents interviewed (19.0%, N = 46) included Parole Officers (N = 25), CPOs/Program Facilitators (N = 10), Parole Officer Supervisors (N = 7), Program Managers (N = 5), Case Management Officers (N = 2), a Psychologist (N = 1), and a Case Management Supervisor (N = 1).

2.3.3 Other Data Sources

Documents reviewed for various components of the evaluation were:

  • CSC National Program Guidelines
  • Standard Operating Practices relevant to correctional programming;
  • Commissioners Directives (CDs) relevant to correctional programming;
  • Corrections and Conditional Release Act (CCRA);
  • Detailed program descriptions provided by Reintegration Programs;
  • Information provided by CSC staff members to the evaluation team;
  • Documents provided by inmates to the evaluation team24; and
  • A review of the relevant literature regarding the relevancy, success, and cost–effectiveness of correctional programs.

 

2.4 Measures

Supplementary Intake Assessment Instruments

Offenders’ programming needs are assessed to determine in which correctional program, and at what intensity level, he or she should participate. Referrals to correctional programs occur on the basis of several assessment activities and tools, including:

  • an interview with the offender;
  • objective tools assessing static and dynamic indicators; and
  • supplementary specialized assessments conducted at intake (e.g., supplementary intake assessments) for programs including substance abuse, family violence, violent offenders, and sex offender.

The following section provides a description of the supplementary intake assessment used by target program.

Substance Abuse

As mandated by Standard Operating Procedure 700–04 (s.34), when an offender meets one or more criteria indicating the possibility of substance abuse problems, he/she must be referred to complete a supplementary substance abuse intake assessment. The results of these assessments are then used to guide referrals to the appropriate intensity level of substance abuse programming. Within CSC, the Computerized Lifestyle Assessment Instrument (CLAI) and its successor, the Computerized Assessment of Substance Abuse (CASA), are used to determine offender referrals for substance abuse programming.

Computerized Lifestyle Assessment Instrument (Robinson, Fabiano, Poporino, Millson, & Graves, 1993)

The CLAI is a computer administered inventory with over 600 items assessing substance abuse among offenders. The inventory examines a variety of factors associated with substance abuse, including physical health, mental health, nutrition, family and social functioning, education/work, criminal behaviour patterns, substance abuse and its relationship to criminal activity, past substance abuse treatment, and readiness for future treatment. The CLAI can be administered in both French and English (Weekes, Vanderburg, & Millson, 1995), has been used with Aboriginal offenders (Vanderburg, Weekes & Millson, 1994), and key components have been tested on women offenders (Lightfoot & Hodgins, 1988). Validity studies have demonstrated strong criterion validity for the CLAI (e.g., Robinson & Millson, 1991; Beal, Weekes, Millson, & Eno, 1997; Weekes, Moser & Langevin, 1997).

Computerized Assessment of Substance Abuse

After consultations with operational staff and an international accreditation panel of experts, the Reintegration Programs Branch of CSC developed a new assessment system to replace the cumbersome CLAI. In 1999, the CASA was developed, field tested, and refined by the Addictions Research Centre. The CASA is a 288–item, bilingual, audio–enhanced instrument that assesses substance abuse in seven domains: (a) alcohol abuse severity, (b) drug abuse severity, (c) patterns of use, (d) link to criminal behaviour, (e) parental substance abuse, (f) previous program participation, and (g) treatment readiness. Research suggests that the CASA accurately differentiates cases for referral to SAPs (Kunic & Grant, 2006). Specifically, offenders with increasing substance abuse severity levels as measured by the CASA had more criminogenic need indicators as measured by the OIA, had more involved criminal histories as measured by higher static factor (risk) scores and were rated more likely to re–offend by the Statistical Information on Recidivism Scale Revised (SIR–R). Due to the time period covered in the present evaluation, scores from either the CLAI or the CASA are used as a measure of need for substance abuse programming.

Family Violence

Spousal Assault Risk Assessment (SARA; Kropp, Hart, Webster & Eaves, 1995)

When screening criteria on the Family Violence Risk Assessment (FVRA) are met, the offenders Case Management Officer will complete the SARA. Results of this assessment will determine whether the offenders level of risk is low, moderate, or high. The SARA has been found to be useful in guiding Case Management Officers in making recommendations regarding family violence programming, private family visits, and for release considerations (Gitzel, 1997). According to these suggestions, offenders classified as low risk on the SARA should be referred to a Family Violence Awareness Program, whereas offenders assessed as moderate or high risk should be referred to Moderate or High Intensity Family Violence Treatment Programs, respectively.

Violent Offenders

Statistical Information on Recidivism (SIR; Nufield, 1982)

The SIR combines 15 items in an actuarial scoring system that provides a probability of re–offending within three years of release. Each of the 15 items is static and includes type of current offence, age of admission, previous incarceration and history of violence. Twelve of the 15 items relate to criminal history, and the remaining three items refer to marital status, number of dependants, and employment status at arrest. Summation of SIR items yields a total score ranging from –30 (poor risk) to 28 (very good risk). The items of the SIR are negatively weighted, with lower numbers indicating greater severity. Prediction studies have shown the SIR to be predictive of general recidivism (Wormith & Goldstone, 1984; Bonta, Harman, Hann & Cormier, 1996; Kroner & Loza, 2001; Mills, Loza & Kroner, 2003; Mills, Kroner & Hemmati, 2004). An offenders SIR assessment is used to guide referral to the appropriate intensity level for VOPs.

Sexual Offenders

Specialized Sex Offender Assessment (SSOA)

SSOAs must be completed for: (a) offenders whose current offence is a sex offence; (b) offenders who have a history of sex offences; and (c) offenders whose current or past offences involved sex offences, whether or not the latter resulted in conviction. The SSOA includes an assessment of history and development of sexual behaviour, sexual preference(s), attitudes, cognitive distortions, social competence, medical history, psychopathology, previous assessment results, prior treatment results, other information relevant to a particular individual offender, score on the Static–99 (Hanson & Thornton, 1999), a structured assessment of dynamic risk factors, using actuarial risk assessment instruments with established reliability and validity (STABLE, Hanson & Harris, 2001), additional assessment instruments that may be administered at the discretion of the psychologist and in conjunction with established community standards for assessment, and a clear evaluation of the need for sex offender treatment. The SSOA forms the basis of a referral to a specific intensity SOP.

2.4.2 Programming Needs Assessment Measure

The above supplementary intake assessment criteria were replicated and, when necessary, approximated for the purposes of this evaluation using automated data. This measure was developed to analyse whether offenders who met the referral criteria for a given correctional program were assigned to that program, and moreover, whether those offenders who met the referral criteria for the program in which they were enrolled performed better on institutional and community outcomes, relative to those offenders who did not meet the referral criteria but were nonetheless enrolled in the program. Data were drawn from available automated sources, including CPIC data OMS. The criteria and data sources used to determine whether offenders met the referral criteria for each program are detailed in Appendix C.

2.4.3 Measures of Pre–/Post-Program Belief and Attitude Changes

Generic Program Performance Measure (GPPM; Stewart, 2005)

The GPPM is a measure of treatment gain, involving a pre– and post–treatment rating. The GPPM is based on a rating system used in a number of programs and clinical situations that allows the program facilitator to systematically assess the progress of a participant on the goals of the intervention. A version of the GPPM used by CSC has been shown to have good psychometric properties indicating that the tool is reliable (Hogue, 1993; 1994). The rating is based on behavioural indicators, such as the quality of the participants participation in role–plays, group exercises, quality of homework, indication of skill development, and application (see Appendix E).

2.4.4 Cost of Maintaining an Offender (COMO)

The calculation of the COMO is based on actual salaries and operating expenditures as reflected in the 2006–07 Public Accounts, including contribution to employee benefit plan but excluding the retroactive payments of salaries pertaining to previous years for newly signed collective agreements. It excludes capital expenses and CORCAN (a special operating agency) disbursements. A reallocation of costs for common services (e.g., personnel, finance, material management) and for other shared services (e.g., food, utilities) between certain adjacent institutions (e.g., Westmorland/Dorchester, Laval complex, Joyceville/Pittsburgh) was performed. In addition, a number of central charges (e.g. Employee benefit plan, Perimetric Intrusion Detection System (PIDS) maintenance contract cost, CORCAN training cost, inmate clothing and officers uniforms and training costs of Correctional Management Learning Centre (CMLC) were reallocated to the various sites.

2.4.5 Cost Efficiency

The efficiency of program usage was assessed by assigning expenditures to various activity measures that were associated with correctional programs. Activity measures quantified the number and type of program entries and exits, and were used to assess a dimension of efficiency by comparing the number of program entries with the number and type of program exits. The total activity measured the frequencies (counts) of entries to and exits from a program within given time frames. Activity measures also reflected the types of program exits (terminations) within the examined time frame. To summarize reporting and interpretation, the evaluation team grouped terminations into the three categories that comprise the ITT group: (a) completions; (b) drop outs; and (c) population management. Completers were those offenders who completed the program, either successfully or unsuccessfully. Drop Outs include offenders who did not complete the program, but were terminated due to factors under their own control. For example, drop outs included those terminating program enrolment due to personal misconduct and those refusing to comply with a recommended program. Population Management includes offenders who did not complete the program, but were terminated through no fault of their own. Examples of specific reasons for population management non–completion include transfer to another facility and release from facility.

Costs were aggregated and then calculated for completions and successful completions on a per inmate basis. Allocations of expenditures were then used to determine the percentage of total program expenditures allocated to completions.

2.4.6 Cost Effectiveness

Cost effectiveness was presented in terms of the cost savings associated with the program and outcome measures, and where possible in terms of dollars invested for the level of outcome. Note that outcome measures were designed to assess program effectiveness. For the purposes of this evaluation, outcome measures used in the cost–effectiveness analyses were early release (institutional outcome) and recidivism (community outcome). The level of outcome was calculated in terms of days saved. Institutional and community maintenance costs derived from COMO were then assigned to the outcome measures. A cost savings (where established) was calculated for each program and outcome measure.

Calculation of the Cost Effectivness of Correctional Programs[D] Calculation of the Cost Effectivness of Correctional Programs (institutional out come and community outcome).[D] Calculation of the Cost Effectivness of Correctional Programs (institutional outcome and community outcome).[D]

2.4.7 Semi–Structured Interviews

Semi–structured interviews were used to facilitate the collection of information and to provide stakeholders the opportunity to identify issues that may not have been considered prior to the evaluation. Participation in interviews was solicited by way of a request to contribute information relevant to the evaluation through telephone interview or face–to–face interviews. A single interview protocol was developed for all of the stakeholder groups, with various sections of the interview being more relevant to different groups. CSC staff member stakeholders included Parole Officers, Program Facilitators, Program Managers, Regional Program Managers, Assistant Wardens Intervention, Managers Assessment and Intervention, Psychologists/Psychotherapists, Parole Officer Supervisors, Elders/Spiritual Advisors, Aboriginal Liaison Officers, Behavioural Counsellors, Program Coordinators, Regional Administrators of Correctional Programs, Assistant Deputy Commissioners, Case Management Officers, Clinical Directors/Clinical Team Leaders, Inmate Pay/Employment Officers, Case Management Supervisors, Correctional Manager Psychiatric Hospitals, Personal Development Clerks, Intensive Intervention Strategy Managers, Program Plan Assistants, Guidance Counsellors, and Executive Directors. Interviews were structured such that they addressed the following evaluation objectives: (a) implementation; (b) continued relevancy; and (c) best practices and lessons learned25. Questions were designed such that responses were open–ended, categorical, dichotomous, or rated along a six–point scale26.

2.5 Institutional and Community Outcomes

For the purposes of this evaluation, effectiveness of correctional programs was measured along a continuum of outcomes. This continuum of correctional results began with outcomes observable in the institution, with the commencement of an offenders federal sentence, and ended with post–release outcomes, including provincial as well as federal re–offending. As such, post release outcomes extended past the end of an offenders federal sentence. The series of outcomes were defined as follows:

2.5.1 Pre–/Post–Correctional Program Treatment Change

Most correctional programs have a pre–/post–program assessment battery to help gauge the anticipated attitude and belief changes that occur as a result of participation in the correctional program. The instruments contained in the assessment battery are matched to treatment targets and provide an assessment of immediate impact of program participation.

2.5.2 Violent Institutional Incidents

Violent institutional incidents included charges for fights, assaults, threats, and disrespect to provoke violence.

2.5.3 Non–violent Institutional Incidents

Non–violent institutional incidents were any institutional charges, excluding charges for violent institutional incidents.

2.5.4 Substance–related Institutional Incidents

Substance–related institutional incidents included charges for possession of illegal substances, dealing in contraband, taking intoxicants, and failing or refusing to provide a urine sample.

2.5.5 Discretionary Release

Discretionary release is considered for all offenders who are being released prior to statutory release, or at WED. For offenders who require more structured monitoring in order to maximize reintegration and public safety, discretionary release allows for proper follow–up to take place. There are two forms of discretionary release. Day parole can be granted after an offender has served six months of a sentence, or six months before full parole, whichever is later, with conditions that require the offender to return daily to a penitentiary or a community–based residential facility. Full parole can be granted after an offender has served one–third of a sentence or seven years, whichever is less, allowing the offender to serve the remainder of the sentence in the community prior to the completion of sentence.

2.5.6 Any Readmission

Any readmission included readmissions for technical revocations (violation of terms of conditional release without re–offence, and revocation without offence), and readmissions for new offences (see definitions below).

2.5.7 Readmission with a New (Federal or Provincial) Offence

Readmissions with a new offence included readmissions to either federal custody or provincial custody. Readmissions to federal custody included: (a) new warrant of committal offences (i.e., new sentence of two years or more); (b) violation of terms of conditional release with an offence; (c) revocation with an offence; and (d) revocation with an outstanding charge. Provincial readmissions included any new sentences past the expiry of the federal sentence that resulted in a return to provincial custody (i.e., a sentence of less than two years).

2.5.8 Readmission with a New (Federal or Provincial) Violent Offence

Readmissions with a new federal or provincial violent offence included readmissions with a new offence, where the new offence was violent as defined by the Schedule I offences listed in the CCRA (1992) 27.

2.5.9 Readmission with a New (Federal or Provincial) Sex Offence

Readmissions with a new federal or provincial sex offence included admissions with a new offence, with the new offence being sexual in nature (i.e., sexual assault)28.

 

2.6 Procedures

Interviews were conducted by the evaluation team either in person or by telephone during the month of April 2008. The interview process included site visits to Springhill, Dorchester, Westmorland, Joliette, Saint–Anne–des–Plaines, Archambault, Millhaven, Collins Bay, Pittsburgh, Bowden, P Sakastw, Edmonton, Pacific, Fraser Valley, and Matsqui Institutions as well as Moncton, Ville–Marie, Ottawa, Edmonton Urban and Vancouver Parole Offices. Key sources at RHQ were asked to participate in a telephone interview. Individuals at the sites visited that were unavailable to participate but were interested in being interviewed, were contacted by telephone and subsequently interviewed. Interviews were approximately 25 to 40 minutes in duration.

Qualitative data generated as a result of the interview process were entered into Microsoft Word and organized by different stakeholder groups. Themes relevant to the evaluation objectives were then generated by evaluation analysts. Similarly, quantitative data were entered into a Microsoft Excel spreadsheet, and then converted and analyzed electronically using a Statistical Analysis System (SAS).

Automated data used to identify and analyse the ITT and comparison groups were drawn from CSC’s OMS (N = 48,601 warrant of committal admissions). Next, to establish success or failure on institutional outcome measures, these data were linked to other OMS data relations (e.g., institutional charges). For community outcome measures, the evaluation extended the follow–up period beyond an offender’s federal sentence to include provincial re–offending in the definition of recidivism. As such, offender OMS records were linked to CPIC data, resulting in over 8 million text–based lines of information. These records were then parsed through a routine programmed in SAS software. The resulting electronic database contained provincial offending, and sentencing information subsequently used in the community outcomes analyses.

Financial information regarding individual correctional programs was obtained from CSC’s Integrated Financial and Materials Management System (IFMMS). Data regarding the cost of maintaining an offender were drawn from CSC’s COMO data in the institution and community tables.

 

2.7 Analyses

The evaluation team conducted qualitative analyses for open–ended interview questions, through which themes were created. Themes were generated for evaluation objectives where appropriate (relevancy, success, cost–effectiveness, implementation issues, and unintended effects). These themes are presented in the Key Findings section.

Quantitative methods were used to profile the samples of offenders, to identify trends, and to compare various characteristics with a matched cohort. Specifically, Chi–square analyses were used for within– and between–group comparisons on categorical variables (static and dynamic levels of intervention), while t–tests were used for similar comparisons on continuous variables (time to discretionary release). Survival analyses29 were used to examine and compare between group differences where data were censored30. Cox Proportional Hazard Models were used as the models of choice for the Survival Analysis, where the Hazard Ratios associated with this model are the effect of an explanatory variable (i.e. the effect of program participation) on the hazard or risk of an event. Outcome measures examined for between–group analyses were31: (a) rate of institutional incidents; (b) positive decisions for discretionary release; (c) length of time out in the community; and (d) returns to federal custody (with and without a new offence). Similarly, within–group analyses comparative/outcome measures were those listed above, plus static and dynamic levels of intervention, criminal history, age at admission, and aggregate sentence.

A multivariable approach to modeling was chosen to statistically control for any underlying differences that may be contributed to program effect other then program participation, such as offenders age, risk, motivation level, etc32. However, when the number of offenders who were assigned to a program but did not participate was too small to form an appropriate sized comparison group a one to one matched cohort was selected using the same variables used in the multivariable approach.

All analyses were disaggregated by non–Aboriginal male, Aboriginal male, and women offenders, where applicable. Potential differences among some important variables between the comparison group and treatment groups were controlled for in all analyses. These variables included age at admission, sentence length, dynamic need, and Custody Rating Scale. In some cases where small sample sizes prevented the use of control variables, every effort was made to find the most appropriate matched comparison group. Any analyses for which this was required will be noted, along with its limitations.

When evaluating the effect of program participation on violent institutional charges, a repeated measures Analysis of Variance was conducted to compare the change in the yearly rate of violent institutional charges before and after program involvement. Artificial program start and end dates were assigned to the comparison group in order to compare the rate of violent institutional charges before and after program delivery. Based on data from program completers, program start dates for the comparison group were derived by calculating the average proportion of incarcerated time an offender would normally wait to begin the program. The end dates for the comparison group were derived by adding the number of days allotted to deliver the program to these start dates. By assigning start and end dates to the comparison group, any natural fluctuation in institutional charges among the general offender population was controlled for.

Logistic regression analysis was used to model the likelihood of receiving a discretionary release and the likelihood of readmission after release. As a result of small numbers of individuals belonging to the Population Management and Drop–out groups, only the ITT group and where data allowed, the Completers group could be analyzed.

Receiver Operating Characteristic (ROC) analyses were used to assess the predictive validity of the supplementary intake assessment tools used in the process of referring offenders to correctional programs. The accuracy of the supplementary tools was tested on how well they separated offenders into recidivists and non–recidivists. Accuracy was measured by the area under the ROC curve.

 

2.8 Limitations

There were limitations that should be noted within the current evaluation. Although the evaluation is comprehensive and was able to answer many meaningful questions, time constraints limited the scope of several aspects of the evaluation. For example, the gap between program referrals and actual enrolment could be investigated through qualitative interviews with staff members across the regions, and did not include the input of offenders.

There was a large volume of records used in the evaluation of correctional programs (N = 48,000) including all offenders who had a Warrant of Committal from 1997 to 2007. To be able to analyse this much data, reliance was made on databases that could easily be queried. Data generated and stored in OMS was primarily used for this evaluation. Other sources of data that could have provided additional information were not considered. It should be noted that Correctional Plans, casework records, and psychological reports were not analysed in the current evaluation.

Correctional programs offered in the community and in institutions often engage community partners. Again due to time constraints, the evaluation did not consult with volunteers or community groups on the functioning of correctional programs. As a mitigation strategy, these community partners will be engaged in a subsequent, more targeted evaluation on community reintegration initiatives.

Many offenders have been identified as having multiple treatment needs and participate in multiple programs. When this occurs, it is often difficult to ascertain and isolate the treatment effect of each program. Some analyses were performed to examine multiple needs and multiple treatment participation. However, this was not done for all programs. When a program was analysed for effectiveness against a comparison group, all results are attributed to participation in that program, regardless of other programs the group may have participated in. Treatment effect with multiple program participation has not been isolated.

Program outcomes for women offenders may be limited for a variety of reasons. The Women Offenders Sex Offenders Programs are delivered on a one–on–one basis. The number of participants was therefore restricted, making it difficult to generate a large enough sample and comparison group for analyses. Similarly, the program structure for the Women Offenders Substance Abuse Programs did not facilitate analyses as offenders within the comparison group could have been exposed to components of other substance abuse programs33. Finally, although the current report examined violence, it was limited to the Spirit of A Warrior Program. The Violence Prevention Program for Women Offenders is relatively new. Analyses were not conducted due to the insufficient number of participants having enrolled to date. It is anticipated that such analyses will be conducted once an adequate number of women offenders have taken part in the program.

CPIC data was used to proxy program criteria and outcomes past the end of an offenders sentence. Multiple data checks have been performed on the CPIC database created for the purpose of this evaluation and its accuracy has been assessed and validated. It was assumed within the evaluation that Parole Officers making decisions about offenders programming needs would have access to this information in a timely manner. However, in some cases may not have been the case if records were delayed.

The evaluation, when measuring financial resources, cost–efficiency and cost–effectiveness, relied on data maintained through CSCs financial management system. There may be a lack of precision in financial reporting for Aboriginal offender correctional programs. This would explain, for instance, the relatively low expenditures on the In Search of Your Warrior Program.

 

3.0 Key Findings

The following results are presented under their respective Evaluation Objectives, namely:
(a)
Relevancy; (b) Implementation; (c) Success (effectiveness and efficiency); (d) Cost Efficiency and Cost Effectivenes; and (e) Unintended Findings.

3.1 Objective 1 Relevancy: The extent to which correctional programs are consistent with departmental and government–wide priorities, and realistically address an actual need.


SUMMARY FINDING 1: There exists a large body of evidence, both national and international, which confirms that targeting specific criminogenic needs is a relevant and effective means of enhancing public safety and reducing re–offending. As such, CSC’s correctional programs are consistent with departmental and government–wide priorities.

 

Government Wide Priorities

Canadian federal departments Strategic Outcomes and Program Activities are grouped into 13 long–term benefits to Canadians referred to as Government of Canada Outcomes in three broad sectors: social, economic, and international. Through two Program Activities Care and Custody and Rehabilitation and Case Management CSC contributes directly and indirectly to the social affairs sector, under the Government of Canada Outcome, Safe and Secure Communities (see Table 3; CSC, 2007).



Table 3: CSC Contributions to Government of Canada Outcome
Government of Canada Outcome CSC’s Direct Contribution CSC’s Indirect Contribution
  Canada’s Social Affairs
   Safe and Secure    Communities
  • Decreased levels of crime and victimization by offenders through the delivery of programs  and services that reduce recidivism.
  • Enhanced community capacity to deliver programs and services that meet the needs of at–risk populations, through partnerships and formal arrangements with the voluntary sector.
  Maintains safe and secure   communities

Departmental Priorities

In 2006–2007, CSC undertook a comprehensive process to identify new priorities. Five priorities were established in response to the changing offender profile, the paramount importance of public safety and the Governments emphasis on crime prevention:

  • Safe transition of eligible offenders into the community;
  • Safety and security for staff and offenders in our institutions;
  • Enhanced capacities to provide effective interventions for First Nations, Mtis and Inuit offenders;
  • Improved capacities to address mental health needs of offenders; and
  • Strengthened management practices (CSC, 2007).

CSCs PAA most saliently addresses the departmental priorities of safe transition of eligible offenders into the community, and enhanced capacities to provide effective interventions for First Nations, Mtis, and Inuit offenders.

To date, there is ample evidence suggesting that presenting offenders with the opportunity to participate in structured interventions facilitates their safe custody and successful reintegration into the community (Andrews et al., 1990; Anstiss, 2003; Dowden & Andrews, 2000; French & Gendreau, 2006; Lipsey, Chapman, & Landenberger, 2001). The following discussion highlights how correctional programs meet government–wide and departmental priorities, and the evidence upon which correctional interventions are based, focusing on: (a) addressing offenders dynamic risk factors; (b) program success; and (c) cost effectiveness.

Addressing Offenders Dynamic Risk Factors

Evidence supporting the relevancy of correctional programs indicates that incarceration alone without interventions does not contribute to successful reintegration (Andrews et al., 1990; Cleland, Pearson, & Lipton, 2006; Lipsey & Cullen, in press). In addition, there is a sound body of empirical evidence that offenders who participate in correctional programs that target their risk and needs have lower rates of institutional misconducts (French & Gendreau, 2006), and are less likely to re–offend following their release, relative to offenders who do not participate in programs (CSC, Offender Programs and Reintegration Branch, 2007). Programs can also be an important vehicle for motivating offenders to change, a desire that can be lacking in this population (Kennedy, 2001; Tellier & Serin, 2001).

Program Success

The definition of program success in the correctional literature varies considerably. The Evaluation Branch of CSC, however, examines success both in terms of program effectiveness the programs ability to meet its stated objectives, and efficiency the extent to which the program is producing its planned outputs in relation to the resources expended for its delivery. Although there is a solid body of empirical evidence attesting to the effectiveness of correctional programs, research documenting the efficiency of such programs is lacking.

One of the most salient indications of a programs effectiveness is its ability to reduce recidivism rates. After more than 30 years of research, there is now ample evidence suggesting that correctional programs are achieving this objective (Andrews et al., 1990; Anstiss, 2003; Dowden & Andrews, 2000; Lipsey et al., 2001). Across studies, effective correctional programs reduce crime by an average of 29% (CSC, Offender Programs and Reintegration Sector, 2007). There is also an abundance of evidence suggesting that the greater the number of dynamic risk factors targeted, the larger the reduction in recidivism rates (Andrews, 2001). Importantly, meta–analytic research34 has demonstrated that the strongest reductions in recidivism occur35 when the risk, need and responsivity principles are applied. That is when: (a) more intensive programs are reserved for higher risk offenders (risk principle); (b) programs target offenders dynamic risk factors (need principle); and (c) interventions are attentive to individuals learning styles and personality (responsivity; Andrews & Bonta, 2003). Furthermore, the applicability of these principles has been established across diverse offender populations, including male offenders (Andrews & Bonta, 2003; Gibbons, 1999), women offenders (Dowden & Andrews, 1999), violent offenders (Dowden & Andrews, 2000), and offenders in both community and institutional settings (Andrews et al., 1990; Dowden, 1998; Hill, Andrews, & Hoge, 1991).

These findings clearly suggest that correctional programs work. Success rates may vary, however, according to the type of program in question. Below is a review of the literature on correctional programs that have been found to work internationally.

Violent Offender Programs

Despite long–standing concerns surrounding the management and rehabilitation of violent offenders, the development and evaluation of VOPs has only recently begun to gain prominence. In general, a common component across these programs is a module focusing on emotions and anger management, and the vast majority of interventions assume a cognitive–behavioural approach (Serin & Preston, 2001). Studies of the VOPs offered in Canada indicate that they are highly effective, revealing a 48% to 86% reduction in violent recidivism (Cortoni, Nunes, & Latendress, 2006; Dowden, Blanchette, & Serin, 1999; Wong, 2001), and a 69% decrease in non–violent recidivism (Dowden et al., 1999). Participation in a VOP is also associated with significantly lower rates of institutional misconducts from pre– to post–program assessment (Cortoni et al., 2006). Although findings from research conducted in other countries are limited by their small sample sizes, they nevertheless provide moderate support for the ability of cognitive–behavioural violence prevention programs in reducing violent recidivism rates and improving survival time in the community before re–incarceration (Berry, 1999; Polaschek, Wilson, Townsend, & Daly, 2005).

Family Violence Programs

Very little attention has been devoted to the effectiveness of FVPP with criminal populations. The few studies that have been undertaken do suggest, however, that recidivism rates can be reduced as a result of such programs. For example, one recent Canadian study reporting a 67% decrease in spousal violence recidivism, and a 50% reduction in general violence recidivism, for program participants compared to non–participating family violence offenders (British Columbia Institute Against Family Violence, 2004, as cited in Motiuk, Cousineau, & Gileno, 2005). Consistent with the literature highlighted previously for the general offender population, cognitive–behavioural approaches appear to be most effective (Saunders, 1996).

Substance Abuse Programs

SAPs for Canadian federal offenders, such as CSC’s NSAP–High, and Intensive Support Units (ISU) for substance abusing offenders, have demonstrated successful results. Recent outcome evaluations of substance abuse interventions in Canada have reported 30% to 50% reductions in recidivism, as well as improvements in substance consumption patterns, for program participants relative to controls (Grant, Kunic, MacPerson, McKeown, & Hanson, 2003; Grant, Varis, & Lefebvre, 2004; Poporino, Robinson, Milson, & Weekes, 2002; T3 Associates, 1999). Research drawn from the U.S. likewise suggests that SAPs lower recidivism rates, though outcomes differ according to the type of intervention used, duration of participation, and whether aftercare is provided (Pearce & Holbrook, 2002). Overall, programs that focus on skill development (i.e., as opposed to non–directive, insight–oriented approaches), emphasize cognitive–behavioural factors, include a structured relapse prevention component, and involve a gradual transition into an extended maintenance program, maximize treatment outcome (Pearce & Holbrook, 2002; Poporino et al., 2002).

Sex Offender Programs

Programs for sex offenders appear to be moderately effective, with the most recent and comprehensive meta–analysis to date revealing a 37% reduction in sexual recidivism and 31% reduction in general recidivism, for program participants relative to comparison groups (Lsel & Schmucker, 2005). Furthermore, the research is clear that only those programs designed specifically for sex offenders have a significant treatment effect and the most effective programs are cognitive–behavioural (Hanson et al., 2002; Lsel & Schmucker, 2005) and are most effective when they adhere to the principles of risk, need, and responsivity (Hanson et al., 2002). These programs reduce recidivism rates by as much as 40% (Hanson et al., 2002; Lsel & Schmucker, 2005).

Life Skills Programs

Originally developed in Canada, and now offered world–wide, the R & R program has been subjected to considerable empirical scrutiny (Pearson, Lipton, Cleland & Yee, 2002; Ross & Ross, 1995). A recent meta–analytic review reveals a 14% reduction in recidivism for program participants relative to controls, and provides support for the programs effectiveness across the spectrum of offender risk levels, in community and institutional settings, and in Canada, the U.S., and the UK (Tong & Farrington, 2006). Lastly, the Counter–Point program offered by CSC and designed to target criminal attitudes and self–management skills, has demonstrated promising results. One recent report found that released offenders who completed the program showed a 32% reduction in suspension rates, a 46% reduction in revocation rates, and a 38% decrease in new offence convictions (Yessine & Kroner, 2004).

Culturally Sensitive Programs

Numerous reviews have highlighted the importance of culturally competent programs for offenders (Zellerer, 2003). In Canada, the advent of programs sensitive to the unique, culture–specific needs of Aboriginal offenders has resulted in significant treatment gains for this population (Proulx & Perrault, 1997; Trevethan, Moore, & Allegri, 2005). VOPs designed for male Aboriginal offenders, for instance, have reported reductions in several need areas, higher reintegration potential ratings, and lower re–admission rates following program completion, including re–admissions for a new violent offence (Trevethan et al., 2005). Preliminary evidence likewise indicates that the VOP designed for Aboriginal women offenders in Canada (SOAW) has fostered positive changes in emotions and anger management, and is generally viewed favourably by both staff and participants (Bell & Flight, in press). Lastly, a program targeting family violence among male Aboriginal offenders (i.e., AHIFVPP) has proven valuable from participants vantage–point for helping them understand and control their violence in intimate relationships (Proulx & Perrault, 1997). In this study, follow–up programming in the community was considered to be an important contributor to mens healing process and their commitment to non–violence (Proulx & Perrault, 1997).

Gender–Specific Programs

The implementation of the Women Offender Program Strategy at CSC (Fortin, 2004) has provided an opportunity for women offenders to benefit from programs that were developed specifically to meet their unique needs and styles of learning. Prior to this initiative, women offenders were offered services that were based on models derived for their male counterparts and, accordingly, early research suggested that treatment for women was inappropriate and ineffective (Blanchette, 2001). Given that the emphasis on gender–informed programming is relatively new, few rigorous outcome evaluations of these programs have been undertaken. The preliminary research that has emerged suggests that gender–responsive programs show more promise (Blanchette, 2001; Pettway, 2006), and recent outcome data indicate that participation in such programs can reduce needs in the targeted areas (Furlong & Grant, 2006; Hall, Prendergast, Wellisch, Patten, & Cao, 2004) and recidivism rates (Hall et al., 2004).

Maintenance Programs

Maintenance programs are available to diverse offender populations and hence, target a variety of offender needs. In essence, these programs seek to increase the probability that positive change will be maintained either following completion of a core treatment program in the institution, and/or upon offenders return to the community. Programs that incorporate relapse prevention or self–regulation components, for instance, have been used with substance–abusing offenders (Peters, 1993), sex offenders (Laws, 1999; Ward & Hudson, 1996), and general offenders (Dowden, Antonowicz, & Andrews, 2003), and focus on teaching an individual how to identify high–risk situations, circumvent habitual destructive coping styles, and enhance feelings of self–efficacy when confronted with these situations (Bakker, Ward, Cryer, & Hudson, 1997; Laws, 1999). This cognitive–behavioural approach to self–management has been found to be moderately successful, in that there is a 15% difference in recidivism rates between participants of programs that incorporate a relapse prevention component into their framework and those that do not (i.e., recidivism rates of 42.5% and 57.5%, respectively; Dowden et al., 2003). Moreover, as the number of relapse prevention components targeted in the program increase the therapeutic potential of the program is enhanced considerably (Dowden et al., 2003).

Cost–Effectiveness

The abundance of evidence establishing the ability of correctional programs to reduce recidivism, technical violations, and re–admission rates, clearly attests to the economic returns associated with providing programs to offenders. However, from the growing body of research examining what works and what does not work in offender rehabilitation, it is clear that not all correctional programs are created equal. Overall, the body of literature that examines the cost–effectiveness of correctional programs indicates positive results for programs aimed at reducing re–offending. The following is a review of recent evaluations of the cost–savings associated with correctional programs.

Several reviews provide evidence that correctional programs are cost–effective relative to no treatment (Andrews, Dowden, & Gendreau, 2000; Aos, Phipps, Barnoski, & Lieb, 1999, 2001; Welsh, 2004; Welsh & Farrington, 2000). The widely publicized Washington State Institute for Public Policy study, initially undertaken in 1999, demonstrated that correctional programs can produce considerable cost–saving to both taxpayers (i.e., reduced costs associated with all levels of the criminal justice system) and victims (i.e., medical and mental health expenses, and property damage and losses; Aos et al., 1999). This investigation demonstrated that, on average, every dollar spent on human–services saves the taxpayer approximately $5.00 and victims $7.00. By comparison, other punishment–oriented interventions and intensive supervision strategies yield substantially lower economic returns, averaging only 50¢ to 75¢ for every dollar allocated to the program. These findings were reinforced in the subsequent cost–benefit analysis of crime reduction tactics undertaken by the Institute in 2001 (Aos et al., 2001).

Although the overall findings for the economic efficiency of correctional programs are positive, their cost–savings may vary according to the type of program delivered and the monetary returns examined (i.e., for victims and/or taxpayers). Below is an overview of the limited published data on the cost–effectiveness of targeted correctional programs.

The literature on the cost–savings associated with SOP has produced conflicting results. The first Washington State meta–analysis revealed only small economic returns for taxpayers (i.e., 25 for every dollar invested; Aos et al., 1999), and their subsequent investigation reported that taxpayers do not even break even with the cost of program delivery (Aos et al., 2001). However, a substantial positive benefit emerged when the benefits to crime victims were included, producing an average savings of $4.13 for every dollar spent on the typical cognitive–behavioural program (Aos et al., 2001). Independent studies have also garnered support for the savings associated with sex offender treatment, with Canadian research revealing savings of 4.4 million dollars in criminal justice costs for every 100 sex offenders treated (Marshall, 1992). Moreover, providing treatment can potentially prevent an additional 50 future victims (Marshall, 1992). Research from Australia similarly suggests that a 14% reduction in sexual recidivism can generate nearly 4 million dollars in economic returns for every 100 child sexual offenders treated (Donato & Shanahan, 1999). As noted previously, several studies have reported higher reductions in recidivism among treated sex offenders and, consequently, greater cost–saving associated with the provision of SOPs may be expected.

An independent evaluation of one of Canadas core SAPs found that it generated approximately $2,000 in annual savings per offender (T3 Associates, 1999), and drug treatment programs offered in the U.S. have generally been found to produce approximately $3.00 in benefits per dollar of cost for taxpayers and crime victims (Aos et al., 2001). An independent study in the U.S. also suggests that participating in an institutional substance abuse program can reduce the length of time incarcerated by 51 days compared to the average control group (McCollister et al., 2003). Based on cost calculations, this means that treatment reduces recidivism at a cost of $80 (USD) per day of incarceration. Moreover, for participants who receive both in–prison treatment and aftercare services, an additional day of incarceration can be avoided at a cost of $51 per day compared to offenders who receive in–prison treatment only (McCollister et al., 2003). Other research has confirmed the net savings associated with providing aftercare for substance abusing offenders (Aos et al., 2001).

Cognitive–behavioural or life–skills treatment programs for the general offender population, such as R & R, are highly cost–effective. The cost of the programs per participant is low and results in approximately $2,400 in net taxpayer–only benefits per participant and about $7,500 in net benefits per participant when savings to crime victims are added into cost calculations through lowered recidivism rates (Aos et al., 2001).

The information published for correctional programs as an aggregate are generally optimistic in the cost–savings associated with existing programs. Quantitative reviews and independent studies have shown that correctional programs are associated with considerable cost–savings for both taxpayers and victims. Effective programs can reduce costs by minimizing the resources expended at all levels of the criminal justice system and the monetary losses associated with being a victim of crime. The intangible costs for victims, such as the pain, suffering, and loss of life that can result from being victimized, have not been quantified and clearly are immeasurable in terms of their value. There are no available data on the cost–effectiveness of programs targeting Aboriginal, women, violent, or family violence offenders. The present evaluation will examine the cost–effectiveness of Aboriginal– and women–specific programs, and will also disaggregate results for non–cultural/gender–specific programs by non–Aboriginal male offenders, Aboriginal male offenders, and women offenders, where possible.

Taken together, the above findings suggest that correctional programs contribute to offender management, rehabilitation, and safe and timely reintegration. As such, correctional programs produce results for Canadians, both in terms of enhancing public safety and reducing the costs associated with the re–arrest, prosecution, and incarceration of offenders. Consequently, the continual evaluation of their success remains instrumental to ensuring that only those programs that receive empirical support remain in operation. In this regard, international interest in what works in corrections has made substantial contributions to contemporary correctional practices.

3.2 Objective 2: Implementation:

This evaluation objective ascertains whether correctional programs are organized or delivered in such a way that goals and objectives can be achieved. This involves appropriate and logical linkages between activities, outputs, outcomes and long–term outcomes.


SUMMARY FINDING 2: The criteria used to refer offenders to specific correctional programs are appropriate, although supplementary assessment tools that are actuarial in nature and that incorporate static risk were shown to discriminate intensity level requirements more accurately.


3.2.1 Referral Criteria

FINDING 1: The criteria used to refer offenders to specific correctional programs are appropriate, as those identified as having a high level of program need as per the target program referral criteria were more likely to re–offend after than those assessed as ‘having no’ or ‘low need’ for programs. The supplementary intake assessments used to assess and designate the required program intensity levels are most accurate when static risk is considered within the assessment.

To assess the predictive validity of CSC’s program referral criteria, referral decisions (i.e., to refer or not refer based on referral criteria outcome) were examined for all offenders released between 1997 and 2007 and available for a three–year follow–up period (N = 23,724). Criteria were deemed to be ‘accurate’ if offenders who were designated as ‘high risk’ and, thus, as having a need for the target program, were more likely to re–offend than those not meeting the referral criteria. In order to avoid confounding the effects of those who participated in programs during incarceration (and thus had their program need area met), the sample was composed of those who did not participate in the target program during their incarceration period. For example, it would be expected that those deemed as ‘requiring a violence prevention program’ as per the referral criteria and not participating in the program while incarcerated, would be less successful in the community than those assessed as not requiring the program. Likewise, those deemed as ‘requiring a higher intensity substance abuse program’ as per the referral criteria and not participating in that program while incarcerated would be less successful in the community than those assessed as requiring a lower intensity level of the program. Overall, analyses revealed that non–participants who met the referral criteria were more likely to return to custody after being released into the community. That is, those offenders who met the referral criteria and did not receive the associated program were more likely to return to custody.

The referral criteria for program participation are based, to varying degrees, on the outcomes of various actuarial tools36. In most cases, the actuarial tool facilitates the designation of program intensity levels (high, moderate or low intensity)37. The following analyses present the accuracy of overall referral criteria and the accuracy of the associated supplementary intake assessment tool, disaggregated for non–Aboriginal men, Aboriginal men and women where appropriate.

 

Violent Offender Program

Referral to a VOP is based largely on static risk level rating and criminal offence history. As such, the analyses of referral criteria examined both referral criteria outcome (referred, not referred) and the actuarial tools used to predict re–offending (SIR–R1 scale and overall static risk level rating).

Analyses revealed that, for both non–Aboriginal and Aboriginal male VOP non–participants, offenders who met the referral criteria for a VOP program were more likely to return to custody with a new offence than those who did not (see Table 4).


Table 4: Community Correctional Outcomes for Violent Offender Program Non Participants
  Non–Aboriginal Men Aboriginal Men
 Community  Correctional  Outcome Met Referral Criteria (%) Did Not Meet Referral Criteria (%) Met Referral Criteria (%) Did Not Meet Referral Criteria (%)
 Return with a new  offence 36.3*** 21.5*** 40.8*** 32.6***
 Return with a new  violent offence 6.7*** 2.9*** 7.5*** 4.1***

Note: ns = not significant; *p < .10, **p < .05, ***p < .01.

With respect to accuracy, the actuarial tools38 used to refer offenders into VOPs demonstrated good predictive validity for non–Aboriginal male offenders, as verified by the Areas Under the Curve (AUC)39. Results were not as robust for Aboriginal male offenders, suggesting the need for a risk prediction tool for this demographic of the population (see Table 5).

 

Table 5: Area Under the Curve for the SIR–RI and the Overall Static Risk Factor Level Rating
  Non–Aboriginal Men Aboriginal Men
 Community Correctional Outcome  AUC for the SIR–R1  AUC for Overall Static Risk Factor Level
 Return with a new offence 0.69*** 0.61***
 Return with a new violent offence 0.63*** 0.57***

Note: ns = not significant; *p < .10, **p < .05, ***p < .01.

 

Family Violence Prevention Program

Analyses revealed that, amongst non–Aboriginal male FVPP non–participants, offenders who met the referral criteria for a FVPP program were more likely to return to custody with a new offence than those who did not (see Table 6). Findings were not statistically significant for Aboriginal men.

 

Table 6: Community Correctional Outcomes for Violent Offender Program Non Participants
  Non–Aboriginal Men Aboriginal Men
 Community  Correctional Outcome Met Referral Criteria (%) Did Not Meet Referral Criteria (%) Met Referral Criteria (%) Did Not Meet Referral Criteria (%)
 Any return 44.6*** 31.2 39.2ns 37.0
 Return with a new offence 29.1* 23.7 40.0ns 36.8
 Return with a new violent offence 3.2 ns 3.1 7.8 ns 4.5

Note: ns = not significant; *p < .10, **p < .05, ***p < .01.

With respect to predictive accuracy, the SARA (Kropp et al., 1994, 1995, 1998), used to refer offenders to FVPP, did not discriminate between recidivists and non–recidivists, as demonstrated by AUCs obtained (see Table 7). In addition, when SARA scale items were recalibrated and an actuarial score established, the resulting overall score did not discriminate between recidivists and non–recidivists40.

 

Table 7: Area Under the Curve for the Spousal Assault Risk Assessment
  Non–Aboriginal Men Aboriginal Men
 Community Correctional Outcome AUC for the SARA AUC for the SARA
 Any Return 0.51* 0.50 ns
 Return with a new offence 0.51 ns 0.51 ns
 Return with a new violent offence 0.51 ns 0.55* ns

Note: ns = not significant; *p < .10, **p < .05, ***p < .01.

 

Substance Abuse Programs

Referrals to SAPs are made on the basis of: (a) substance abuse need outcomes, as per the supplementary intake assessment tool used; and (b) an established link between crime and substance abuse. As such, both referral criteria outcomes and the supplementary intake assessment tools used to measure substance abuse need (the CASA and the Overall Dynamic Substance Abuse Need Level rating as per the OIA) were examined.

The CASA was found to be more accurate in identifying programming needs overall when compared to the Substance Abuse Need Level indicator. Specifically, amongst those who did not receive a SAP while incarcerated, those identified as having a substance abuse programming need as per the CASA were more likely to return to custody (see Table 8).

 

Table 8: Community Correctional Outcomes for Those Meeting Substance Abuse Program Referral Criteria as per the CASA: Non–Participants
  Non–Aboriginal Men Aboriginal Men Women
 Community Correctional Outcome Met Referral Criteria Did Not Meet Referral Criteria Met Referral Criteria Did Not Meet Referral Criteria Met Referral Criteria Did Not Meet Referral Criteria
 Any return 35.7*** 18.6*** 37.9*** 26.0*** 33.5*** 13.5***
 Return with a new offence 27.7*** 19.0*** 37.3*** 29.5*** 18.6*** 7.6***
 Return with a new violent offence 4.1*** 3.1*** 5.5 ns 4.9 ns 1.6 ns 0.5 ns

Note: ns = not significant; *p < .10, **p < .05, ***p < .01.

Next, the overall Substance Abuse Need Level indicator was more accurate in assessing appropriate program intensity levels, suggesting that substance abuse programming needs be assessed based on a combination of CASA and overall Substance Abuse Need Level indicator level outcomes (see Table 9).

 

Table 9: Areas Under the Curve for Community Correctional Outcomes: CASA versus the Overall Substance Abuse Need Level Rating: Non–Participants
  Non–Aboriginal Men
(AUC)
Aboriginal Men
(AUC)
Women
(AUC)
  Community  Correctional  Outcome CASA Overall Substance Abuse Need Level CASA Overall Substance Abuse Need Level CASA Overall Substance Abuse Need Level
 Any Return 0.59*** 0.64*** 0.56*** 0.56*** 0.68*** 0.70***
 Return with a new offence 0.53*** 0.57*** 0.50 ns 0.53** 0.56** 0.61***
 Return with a new violent offence 0.56*** 0.58*** –– 0.55*** 0.64 ns 0.64*

Note: ns = not significant; *p < .10, **p < .05, ***p < .01.

 

Sex Offender Programs

Referral to SOPs is based on a combination of static risk level, dynamic risk level and specialized assessments. For the purposes of this evaluation, programming needs were approximated using a priority rating based on the actuarial tools currently in use to predict re–offending (i.e., the Static 99 and the Stable 2000).

Analyses revealed that, non–Aboriginal male SOP non–participants who met that programs approximated referral criteria were more likely to return to custody with a new sex offence than those who did not (see Table 10). Results indicate a similar trend for Aboriginal male offenders, although results were not statistically significant.

 

Table 10: Community Correctional Outcomes for Those Who Met the Approximated Sex Offender Program Referral Criteria: Non–Participants
  Non–Aboriginal Men Aboriginal Men
 Community Correctional Outcome  Met Referral Criteria  Did Not Meet Referral Criteria  Met Referral Criteria  Did Not Meet Referral Criteria
 Any Return 28.9 31.4* 31.0 33.5***
 Return with a new offence 22.6 31.4* 31.0ns 33.5***
 Return with a new violent offence 6.0 3.8*** 7.0 ns 5.9 ns
 Return with a new sex offence 0.8 0.15*** 0.4 0.1 ns

Note: ns = not significant; *p < .10, **p < .05, ***p < .01.

 

The priority rating outcomes41 used to refer offenders to SOPs demonstrated good predictive validity for non–Aboriginal male offenders, as indicated by the AUCs for sexual re–offending (see Table 11).

Table 11: Area Under the Curve for the Sex Offender Priority Rating
  Non–Aboriginal Men Aboriginal Men
 Community Correctional Outcome  AUC for the Priority Rating  AUC for the Priority Rating
 Any Return  0.64***  0.61***
 Return with a new offence  0.62**  0.57**
 Return with a new violent  offence  0.64 ***  0.54 ns
 Return with a new sex  offence  0.62 ns ––

Note: ns = not significant; *p < .10, **p < .05, ***p < .01, –– dats were not used appropriate for the model used.

 

3.2.2 Supplementary Intake Assessment Tools

FINDING SUMMARY 2: Overall, staff members reported that the supplementary intake assessment tools are useful when prioritizing offenders’s correctional programming needs.

 

Overall, staff members reported that supplementary intake assessment tols, including the CLAI/CASA, the FVRA/SARA, and the SSOA, are used when priotitizing offenders’ correctional plans. Of note, staff from the Prairie and Ontario Regions rated the usefulness of the CLAI/CASA lower than staff elsewhere, while staff from multi–level security institutions rated the usefulness of this tool higher than staff from all other security levels. On average, staff rated their agreement as to whether the CLAI/CASA is used in decision–making lower than that for both the FVRA/SARA and the SSOA. Reasons cited for not finding supplementary intake assessments useful most commonly included problems resulting from offender self–report methodologies, subjectivity, and a need for additional information.

The majority of respondents stated that they find the CLAI/CASA (71.4%, N = 132), the FVRA/SARA (83.3%, N = 130), and the SSOA (87.1%, N = 108) to be useful when prioritizing offenders’ programming needs. Two–way contingency table analysis of responses indicated that staff from the Prairie and Ontario Regions were significantly more likely to indicate they did not find the CLAI/CASA useful, whereas those from the Pacific and Atlantic Regions were more likely to state they find this assessment tool useful, (x2(4) = 22.34, p < .0001). In a two–way contingency table analysis of security level and agreement, staff from multi–level security institutions were more likely to find the CLAI/CASA useful than staff from all other security levels, parole offices, or regional offices. There were no significant differences between regions (x2(4) = 4.00, p = .41) or security levels (x 2(5) = 4.77, p = .45) regarding the usefulness of the FVRA/SARA. However, staff from the Prairie Region were more likely to indicate that they did not find the SSOA useful, whereas those from the Atlantic Region were more likely to indicate that they find this assessment instrument useful (x 2(4) = 11.74, p < .05). Staff from different security levels did not differ in their responses regarding the usefulness of SSOA in prioritizing offenders’ correctional programming needs (x 2(5) = 6.78, p = .24).

To follow up, staff respondents were asked to provide reasons as to why they did not find the CLAI/CASA, FVRA/SARA, and SSOA useful in prioritizing programming needs. Of the 51 staff members who provided reasons regarding the CLAI/CASA, the majority of respondents (66.7%, N = 34) indicated that the self–report format of the CLAI/CASA was problematic. Other reasons included that it should be used in conjunction with additional information or assessment tools (23.5%, N = 12), that the CLAI/CASA were not always completed or were not completed in a timely fashion (7.8%, N = 4), and that the CLAI/CASA often under–classified high need offenders as moderate or low need (7.8%, N = 4).

Of the 25 staff members who provided reasons for not finding the FVRA/SARA useful, many (48.0%, N = 12) stated that it is too subjective or that the results were often overridden based on subjective judgment. Other reasons provided for not finding this tool useful included that it should be used in conjunction with additional information or other assessment tools (24.0%, N = 6), that more staff training is required for the administration and/or interpretation of the FVRA/SARA results (20%, N = 5), and that the information provided by the FVRA/SARA is not useful (16%, N = 4). Fewer respondents indicated that the FVRA/SARA is not culturally sensitive to Aboriginal offenders (8.0%, N = 2), and that the FVRA/SARA should be used with women offenders (8.0%, N = 2).

Of the 15 respondents who provided reasons for not finding the SSOA useful, some indicated that the assessment was not always completed or was not completed in a timely fashion (20.0%, N = 3), that the self–report format of the tool is problematic (13.3%, N = 2), that the information provided by the SSOA is not useful (13.3%, N = 2), and that the tool should be replaced by or complemented with an in–person assessment (13.3%, N = 2) or used in conjunction with additional information or another assessment tool (13.3%, N = 2). A comprehensive list of responses for each of the assessment tools is presented in Appendix H.

Overall, the majority of respondents agreed that the CLAI/CASA, FVRA/SARA, and the SSOA are used in decision–making, such as in the development of an offenders correctional plan or by the CIB. Staff respondents were asked to indicate their agreement on a scale from 1 to 6, with 1 being Strongly Disagree and 6 being Strongly Agree. Distributions and average ratings are presented in Table 12.

 

Table 12: Ratings of Staff Agreement Regarding the Use of Supplementary Assessment Tools in Decision–making
  Degree of Agreement % (N)    
   Strongly  Disagree/Disagree   Somewhat  Disagree  Somewhat  Agree  Agree/Strongly  Agree  M  SD
 Assessment  Tool
  CLAI/CASA 2.7 (5) 7.0 (13) 15.1 (28) 75.2 (140) 5.02 1.09
  FVRA/SARA 6.2 (12) 11.8 (23) 21.5 (42) 60.5 (118) 5.18 1.01
  SSOA 2.5 (3) 1.6 (2) 7.4 (9) 88.5 (108) 5.43 0.96

There were no significant regional differences in mean ratings of agreement between regions for the CLAI/CASA (F(4,179) = 1.79, p = .13) or the SSOA (F(4,116) = 0.94, p = .44). Regional differences for the FVRA/SARA approached significance (F(4,147) = 2.36, p = .06), but post–hoc comparisons were non–significant after applying a Bonferroni correction to account for the number of comparisons. Staff from maximum security institutions rated their agreement significantly lower than those from minimum security institutions (Mean difference = 1.08, p < .05) and Parole Offices (Mean difference = 1.06, p < .05); F(5,176) = 2.77, p < .05). On average, respondents differed in their respective ratings of agreement for each assessment tool, F(2,216) = 8.2, p < .0001. Overall, staff rated their agreement as to whether the CLAI/CASA is used in decision–making lower than they did for both the FVRA/SARA (t(114) = –2.06, p < .05) and the SSOA (t(112) = –4.07, p < .0001). Respondents also rated their agreement as to whether the SSOA was used in decision–making higher than they did the FVRA/SARA (t(114) = –2.55, p < .05).

 

3.2.3 Staff Recommendations

FINDING SUMMARY 3: The majority of staff who had recommendations for improving the usefulness and/or efficiency of supplementary intake assessments suggested using assessments in conjunction with additional information, improving staff training for those administering and interpreting assessments, and allowing for overrides of assessment results with clinical judgment.

 

When asked how the usefulness and/or efficiency of supplementary intake assessments might be improved, the majority of those who responded (N = 137) suggested that assessments should be used in conjunction with additional information (N = 40, 29.2%), and that the objectivity and/or validity of assessments should be improved (N = 30, 21.9%). Additional suggestions included improving staff training and/or ensuring assessors are adequately trained (N = 24, 17.5%), allowing for overrides of assessment results with clinical judgment (N = 23, 16.8%), and developing new assessment tools and/or expanding upon existing tools (N = 19, 13.9%). A complete list of suggestions provided by staff is provided in Tables in Appendix I.

FINDING SUMMARY 4: The majority of staff interviewed received training in at least one program area, and subsequently delivered the program(s) in which they were trained. This was especially prnounced among staff from the Prairie and Atlantic Regions, and less so for other regions, where staff who received training were less likely to deliver all of the programs in which they were trained. The majority of staff who had been trained in substance abuse, family violence, violence prevention, and/or sex offender programs felt that their training spent enough time teaching the knowledge and skills they require in order to deliver programs, that the training is delivered efficiently, and that the training covers the appropriate program content.

 

The majority of staff respondents (55.3%) indicated that they had been trained to deliver at least one correctional program (N = 131). Many staff members reported having been trained in National SAPs, followed by living skills programs such as R & R / Cognitive Skills and A & EM. The proportion of respondents trained in each programs area is listed in Appendix I, Table I10.

Of those staff members who were trained to deliver a program, the majority subsequently delivered at least some (28.7%, N = 37), if not all (64.3%, N = 83) of the programs in which they were trained. For those who did not deliver all or any of the programs in which they were trained (7.0%, N = 9), reasons for not delivering the program(s) included that the program was not offered (21.2%, N = 7), they changed jobs or institutions (9.1%, N = 3), they made a personal decision to not deliver the program (9.1%, N = 3), the program was assigned to another facilitator (9.1%, N = 3), their training was used for purposes other than to prepare for program delivery (9.1%, N = 3), and that they were newly trained and had not yet had the opportunity to deliver the program (9.1%, N = 3). Additional reasons for not delivering programs are listed in Appendix I, Table I11.

Relative to staff from other regions, staff members from the Pacific (SR = 2.0) and Ontario Regions (SR = 2.1) were significantly more likely to indicate that they had not delivered the programs in which they had been trained (χ2(8) = 32.89, p < .0001). Staff from the Pacific Region were also most likely to indicate they had delivered some of the programs in which they had been trained (SR = 2.8), whereas staff from the Prairie (SR = 3.0) and Atlantic Regions (SR = 2.4) were most likely to indicate they had delivered all of the programs in which they had been trained. Across security levels, staff from minimum security institutions were most likely to indicate they had delivered all of the programs in which they had been trained to deliver (SR = 2.3), and regional offices were most likely to indicate they had delivered some of the programs in which they were trained (SR = 3.2; χ2(10) = 25.76, p < .005). However, results from these regional and security level two–way contingency table analyses must be interpreted with caution due to the low number of respondents who reported not delivering any of the programs in which they had been trained (7.0%, N = 9).

The majority of staff who had been trained in substance abuse, family violence, violence prevention, and/or sex offender programs felt that their training spent enough time teaching the knowledge and skills they require in order to deliver programs, that the training is delivered efficiently, and that the training covers the appropriate program content. Endorsements for each program area and indicator of training are displayed in Table 13. Due to the low number of staff who responded No, meaningful contingency analyses could not be conducted between program areas or across regions and security levels.

Table 13: Staff Respondents Who Said “Yes” to Various Indicators of Programs Taining
    Subtance Abuse
N=83
 Family Violence
N=29
 Violent Offender
N=28
 Sex Offender
N=30
 Spends enough time teaching knowledge required to deliver program  67 (80.7%)  20 (69.0%)  19 (67.9%)  17 (56.7%)
 Spends enough time teaching skills required to deliver program  62 (75.6%)  20 (69.0%)  18 (64.3%)  14 (48.3%)
 Training is delivered in an efficient manner  72 (87.8%)  23 (79.3%)  20 (74.1%)  20 (66.7%)
 Training covers the appropriate  program content  68 (81.9%)  23 (79.3%)  20 (74.1%)  20 (66.7%)

Few staff (7.6%, N = 17) indicated that they had delivered a program in which they had not received the training to deliver. When asked to indicate why they delivered a program in which they had not received training, respondents provided the reasons listed in Table 14.

Table 14: Reasons Listed for Delivering a Program in Which Staff Members Were Not Trained to Deliver
 Reasons:  Staff Respondents N (%)
 Operational demands  4 (25.0 %)
 On–the–job training  3 (18.8 %)
 Filling in for another CPO  2 (12.5 %)
 Training not available  2 (12.5 %)
 Trained to deliver similar program  1 (6.3%)
 Asked to co–facilitate with trained CPO  1 (6,3%)
 Don’t know/Not sure  1 (6,3%)

Note: Two respondents (12.5%) provided suggestions unrelated to the question.

FINDING SUMMARY 5: Program stakeholders indicated there are insufficient numbers of trained CPOs to deliver programs given the institutional or regional demand. Possible implications of this insufficiency include releasing offenders without institutional interventions, increasing waitlists for programs, delays in release dates for offenders, and an inability of institutions to address offenders risk for recidivism.

 

The majority of staff respondents (79.9%, N = 175) indicated that there are not sufficient numbers of trained CPOs to deliver programs, given their institutional or regional demand. There were no regional differences in this opinion (X3(4) = 2.14, p = .71), although staff from Community Parole Offices were more likely than staff from institutions or regional offices to indicate that there are insufficient numbers of trained CPOs (SR = 2.3; χ2(5) = 10.69, p = .06). When asked about the impact of not having sufficient numbers of trained CPOs, staff most frequently indicated that offenders are released without programs/interventions (26.7%, N = 24), there are long waitlists for programs (16.7%, N = 15), offenders release / parole is delayed (16.7%, N = 15), CPOs are burned out / overworked (15.6%, N = 14), institutions cannot meet offenders needs / fulfill correctional plans (13.3%, N = 12), and the institutions cannot address recidivism rates or risk (13.3%, N = 12).

Despite the positive ratings of CPO training across program areas, many staff respondents (79.3%, N = 107) felt that the efficiency of training could be improved. Of those who provided suggestions (N = 113), most suggested providing an orientation or generic skills training (23.9%, N = 27), offering ongoing training, refresher courses, and/or mentoring (15.9%, N = 18), offering longer training sessions (14.2%, N = 16), offering training more frequently (9.7%, N = 11), providing more training locations across the country (8.0%, N = 9), and offering more training on dealing with offenders in groups / responsivity issues (8.0%, N = 9). A comprehensive list of suggestions to improve the efficiency of CPO training is presented in Appendix I, Table I12.

FINDING SUMMARY 6: The influence of pre–program assessments on subsequent program delivery is inconsistent across program delivery personnel, with half of staff indicating that the results of these assessments influence program delivery, and half indicating no effect. While staff generally found the pre–/post–program assessments useful when writing their final reports, the majority found that the assessments could be streamlined and made more efficient.

 

To gauge the efficiency of the pre–post program assessments, staff were asked a number of questions relating to how they use they results of the assessments, whether or not the assessments are useful to them, and how the pre–/post–assessment process could be streamlined.

First, staff were asked if the results of an offenders pre–program assessment influences the way in which they deliver their correctional program. Results were split almost equally among staff who answered this question (N = 82), with 52.4% (N = 43) of staff indicating that the results of the pre–program assessment did not have any influence over subsequent program delivery, and 47.6% (N = 39) of staff indicating that pre–program assessment results did have an influence over program delivery. Two–way contingency table analyses examining differences in pre–program assessment influence (yes or no) between regions or security levels found no significant differences. Those staff members who provided explanations for altering their delivery approach after the pre–program assessments (N = 40) noted that the assessments most commonly influenced their development of treatment targets (40.0%, N = 16), their identification of responsivity issues (32.5%, N = 13), and their identification of risk/need levels (20.0%, N = 8). Second, staff were asked to rate on a scale from one (1) to six (6), with one representing not at all useful and six representing very useful, how useful pre–/post–program assessment results are when drafting offenders’ final program reports. Of those who responded to the question (N = 81), the greatest proportion felt that pre–/post–assessment results were useful/very useful (43.2%, N = 35) or at least somewhat useful (23.5%, N = 19) when drafting final program reports. Only 7.4% (N = 6) indicated that pre–/post assessment results were somewhat useless, and 25.9% (N = 21) felt that pre–/post–assessment results were not at all or not useful (M = 3.86, SD = 1.68). ANOVAs conducted to determine if any differences existed in mean ratings of usefulness among regions and security levels found no significant differences.

Similarly, staff were asked to rate on a scale from one (1) to five (5), with one representing “never” and five representing “always”, how regularly they refer to an offenders pre–/post–program assessment results when writing their final program reports. Of those who responded to this item (N = 79), the majority referred to the assessments either often or always (60.8%, N = 48), 17.7% (N = 14) referred to the assessments only sometimes, and 21.5% (N = 17) rarely or never referred to the pre–/post–assessments when writing their final program reports (M = 3.67, SD = 1.41). Once again, ANOVAs conducted to determine if any differences existed in mean ratings of referral regularity among regions and security levels found no significant differences.

To investigate why the pre–/post–program assessments were not being used when staff were writing their final reports, those staff who answered “rarely” or “never” to the referral question were asked to detail why this was the case. The most commonly cited reason for not using the assessments was that the assessments are not a good or complete source of information (27.8%, N = 5), that they are difficult to interpret (12.5%, N = 3), that they are not interpreted by a trained professional (12.5%, N = 3), and because the assessment results are easily skewed by the offender (12.5%, N = 3).

Finally, staff were asked whether or not they felt that the offender self–reported pre–/post–program assessment process could be streamlined or made more efficient. Of the 75 staff members who answered this question, 72.0% (N = 54) felt that the process could be made more efficient. Among the most common recommendations to improve efficiency was the need for a decrease in the number of assessment measures (27.3%, N = 15) and the elimination of repetitiveness in assessment measures (20.0%, N = 11). A comprehensive list of all recommendations offered can be found in Table 15.


Table 15: Staff Recommendations to Improve Efficiency of Pre–/Post–Assessment Process
 Recommendations:  Staff Respondents N (%)
 Fewer assessment measures  15 (27.3%)
 Eliminate repetitiveness  11 (20.0%)
 Fewer questions on assessments  7 (12.7%)
 Simplify language  7 (12.7%)
 Offenders could complete assessment measures on  computers  5 (9.1%)
 Pre–post assessment batteries should be eliminated  5 (9.1%)
 Limit assessments to those relevant to  program/offender  5 (9.1%)
 Fewer response options  3 (5.5%)
 Better training for staff  2 (3.6%)
 More assessment measures  1 (1.8%)
 Clarify process  1 (1.8%)
 Needs to be longer  1 (1.8%)
 Use Paulhus deception scale (BID–R)  1 (1.8%)
 Have admin input data, not CPO  1 (1.8%)
 Change when assessments are  administered/delivered  1 (1.8%)
 Research/validate process  1 (1.8%)

Note: 3.6% (N = 2) of staff answered “Dont Know/Not Sure” for the recommendation item. 5.5% of staff recommendations (N = 3) did not reflect the topic area in question.

 

SUMMARY FINDING 3: For all programs with the exception of sex offender programs, there was a high proportion of offenders who met the referral criteria for a program who were not referred, and a proportion of offenders referred to a program who did not participate. There was also a proportion of offenders enrolled in programs for which they had not met the referral criteria. The large pool of candidates who met the referral criteria for a violence prevention program or high intensity substance abuse program was more likely to be serving shorter sentences in a maximum security institution. This group was also typically referred to the moderate intensity substance abuse programs.

 

The following section reports, by program area, the proportion of offenders who met the approximated referral criteria for a correctional program and were not referred to the target program, or were referred and did not enrol. Typically, these offenders were serving shorter sentences and were more likely to have a maximum security designation than their comparison group. The majority also had substance abuse needs identified as per the CLAI/CASA.

Staff members provided insight regarding these findings. When asked to explain the of reasons why an offender might not participate in a correctional program that is included in his or her correctional plan, the majority of the 234 staff respondents mentioned an offenders refusal to participate (65.4%, N = 153), followed by a short sentence length and/or lack of time to complete a program (36.3%, N = 85), a lack of programs available at the institution (35.9%, N = 84), specific responsivity factors (mental illness, cognitive impairment, low educational attainment, etc.; 22.2%, N = 52), the timing of program delivery at the institution (19.2%, N = 45) and long waitlists for programs (18.4%, N = 43). A complete list of reasons described by staff is provided in Appendix I.

When asked to indicated the reasons for which an offender might participate in a correctional program that is not included in his or her correctional plan, staff most frequently mentioned that additional programming needs might be identified after intake (47.2%, N = 110), that the offender expressed an interest in the program (36.1%, N = 84), that the offender wanted to accelerate their release or transfer to lower security (18.5%, N = 43), that the offender was motivated to change his or her behaviour (10.3%, N = 24), that a program included in the offenders plan was not offered, and therefore a substitution is made (9.4%, N = 22), or that the referral criteria to a program changed subsequent to the development of the offender’s correctional plan (6.4%, N = 15).

 

3.2.4 Program Referrals and Enrolments

FINDING SUMMARY 7: Overall, staff members indicated that program referrals are reserved for those with greater risk and need levels, and that the intensity level of the programs delivered matches the offender risk/need profiles. However, staff noted the difficulty in matching intensity levels of programs with risk/need profiles when required programs are not offered at offenders institutions.

 

Staff were asked to rate on a scale from one (1) to six (6), with one representing “strongly disagree” and six representing “strongly agree”, the extent to which referrals to correctional programs are reserved for those offenders with greater risk and needs. Of those who responded to the question (N = 227), the majority agreed or strongly agreed that referrals were reserved for those with the greatest risk/need (50.7%, N = 115) or at least somewhat agreed that this was the case (17.6%, N = 40). Only 16.3% (N = 37) disagreed or strongly disagreed that program referrals are reserved for those offenders with greater risk/need (M = 4.15, SD = 1.45).

ANOVAs conducted to determine if any differences existed in mean ratings of agreement among regions and security levels found no significant differences between groups.

Similarly, staff were asked to indicate whether or not they felt that there was a match between offenders’ risk/need levels and the intensity level of the program delivered. Of those who responded to the question (N = 225), 81.8% indicated that they believed there was a match between offender risk/need level and the intensity of the program that was delivered. Two–way contingency table analyses examining differences in the match between risk/need and intensity level between regions or security levels found no significant differences.

When asked to expand on the reasons for which they believed there was a disparity between offender risk/need levels and the intensity level of the program delivered, 30.6% of staff who responded (N = 19) cited the lack of availability of the program at the required intensity level as the cause. Other causes for disparity included sentence length/timing restrictions (16.1%, N = 10), incongruent/inaccurate assessment results (12.9%, N = 8), and overrides of decisions due to the consideration of other factors (11.3%, N = 7).

 

3.2.5 Violent Offender Program Referrals and Enrolments

FINDING SUMMARY 8: There was a large pool of candidates who met the referral criteria for a specific violent offender program who were either not referred to the program, or who were referred and did not enrol in the program. These offenders were serving shorter sentences and were more likely to have a maximum security designation than their comparison groups. The majority also had substance abuse needs identified. Further, there was a smaller group of male offenders enrolled in violent offender programs despite not meeting all components of the referral criteria.

 

Violent Offender Programs: Violence Prevention Program and In Search of Your Warrior

Non–Aboriginal Male Offenders

From a sample of 29,85642 non–Aboriginal male offender admissions, it was determined that 23.8% (N = 7,100) met the referral criteria for a specific VOP43, namely the VPP or the Aggressive Behaviour Control (ABC) program. The differences between regions of the proportion of non–Aboriginal male offenders that were referred to a VOP but not enrolled is seen in Figure 4. Of those offenders who met the referral criteria for these programs, 15.7% (N = 1,117) were referred to the program, whereas 84.3% (N = 5,983) were not referred. A comparison of referrals to non–referrals within this eligible group revealed non–referred offenders as more likely to have shorter sentences than those referred (M = 3.1, SD = 1.79 years, versus M = 3.9, SD = 2.66 years, respectively; t(6,983) = –14.64; p < .001). An examination of eligible non–referred offenders revealed that many (67.9%, N = 4,063) were referred to a moderate or low intensity substance abuse program and/or living skills program (61.8%, N = 3,697).

Of those who were referred to VPP (N = 1,793), 36.8% enrolled in the program (N = 659). A comparison of the non–enrolments to enrolments revealed that, similar to the eligible non–referrals, this group was more likely to have shorter sentences (M = 4.2, SD = 2.99 years versus M = 5.0, SD = 3.23 years respectively; t(1642) = –5.84, p < .001) and many (70.1%, N = 465) were referred to a moderate or low intensity substance abuse program. The non–enrolled group were more likely to have a maximum security level designation when compared to their enrolled counterparts (30.0% vs. 23.1% respectively; x2(2) = 14.9, p < .01). Among those offenders who were referred to VPP, 37.7% (N = 676) were referred to the program despite not meeting the referral criteria44. The majority of this group met at least some of the criteria (e.g., a rating of poor or very poor on the SIR, but fewer than two convictions for a Schedule I offence). Two–way contingency table analyses revealed that 34.7% (N = 230) were scored as poor or very poor risk on the SIR; of these, 43.5% (N = 100) had prior convictions for two or more Schedule I offences, and another 4.3% had been convicted for homicide. The remaining 68.6% (N = 412) of offenders who were referred to VPP despite not meeting all of the criteria were scored as moderate, good, or very good risk on the SIR. However, of these offenders, 63.8% (N = 263) had been convicted of at least two Schedule I offences, and more than one–third (35.2%; N = 145) had been convicted of homicide. Finally, analyses suggest that there was a group of offenders deemed suitable for VPP given their violent histories. Specifically, within this group, 140 offenders met the referral criteria for the National High Intensity Family Violence Prevention Program (HIFVPP), and 7 offenders met the referral criteria for a SOP.

Figure 4: Proportion of Offenders Referred to Violent Offender Programs and Not Enrolled, Non–Aboriginal Male

From a sample of 29,856 non-Aboriginal male offender admissions
(Click to enlarge)

[D]

Aboriginal Male Offenders

From a sample of 5,73545 Aboriginal male offender admissions, it was determined that 43% (N = 2,461) met the referral criteria for the In Search of Your Warrior (ISOYW) program46. The differences between regions of the proportion of Aboriginal male offenders that were referred to a VOP but not enrolled is seen in Figure 5. Of those offenders who met the referral criteria, 32.1% (N = 791) were referred to the program, whereas 67.9% (N = 1,670) were not referred. A comparison of referrals to non–referrals within this eligible group revealed non–referred offenders as more likely to have shorter sentences (M = 3.18, SD = 1.83 years versus M = 4.03, SD = 2.25 years respectively; t(2,377) = –9.74, p < .001). An examination of eligible non–referred offenders revealed that many (79.0%, N = 1,319) were referred to a moderate or low intensity SAP and/or a living skills program (55.7%, N = 930).

Of those who were referred to the ISOYW program (N = 1,238), 51.5% enrolled in the program (N = 638). A comparison of non–enrolments to enrolments revealed that, similar to eligible non–referrals, this group was more likely to have shorter sentences (M = 3.85, SD = 2.47 years versus M = 4.29, SD = 2.49 years, respectively; t(1,168) = –3.03, p < .001) and many (83.0%, N = 498) were referred to a moderate or low intensity substance abuse program. The non–enrolled group was more likely to have a maximum security level designation when compared to their enrolled counterparts (23.0% vs. 13.8% respectively; x2(2) = 26.10, p < .01).

Among those offenders who were referred to ISOYW, 36.1% (N = 447) were referred to the program despite not meeting the referral criteria47. The majority of this group met at least some of the criteria (e.g., a rating of High Risk, but fewer than two convictions for a Schedule I offence). Two–way contingency table analyses revealed that 63.1% (N = 280) were scored as high risk to re–offend; of these, 78.1% (N = 203) had prior convictions for two or more Schedule I offences, and another 28.5% (N = 74) had been convicted of homicide. The remaining 36.9% (N = 164) of offenders who were referred to ISOYW despite not meeting all of the criteria were scored as moderate or low risk to re–offend; however, of these offenders, 63.0% (N = 97) had been convicted of at least two Schedule I offences, and 20.8% (N = 32) had been convicted of homicide. Finally, analyses suggest that there was a group of offenders deemed suitable for ISOYW given their violent histories. Specifically, within this group, 134 offenders met the referral criteria for the HIFVPP, and 19 offenders met the referral criteria for SOPs.

Figure 5: Proportion of Offenders Referred to Violent Offender Programs and Not Enrolled, Aboriginal Male

From a sample of 5,735 Aboriginal male offender admissions
(Click to enlarge)

[D]

Violent Offender Programs: Spirit of A Warrior

Women Offenders

From a sample of 61848 Aboriginal women offender admissions, it was determined that 14.1% (N = 87) met the referral criteria for SOAW49. The differences between regions of the proportion of women offenders that were referred to SOAW but not enrolled is seen in Figure 6. Of these offenders who met the referral criteria, 43.7% (N = 38) were referred to the program, whereas 56.3% (N = 49) were not referred. A comparison of referrals to non–referrals within this eligible group revealed non–referred offenders as more likely to have shorter sentences (M = 2.61, SD = 0.99 years versus M = 3.75, SD = 2.40 years, respectively; t(82) = –2.95, p < .01). An examination of eligible non–referred offenders revealed that many (77.6%, N = 38) were referred to a moderate or low intensity substance abuse program.

Of those who were referred to the SOAW program (N = 120), 66.7% enrolled in the program (N = 80). A comparison of the non–enrolments to enrolments revealed that, unlike the eligible non–referrals, there were no differences in sentence length between groups. Almost all of the non–enrolments (97.5%, N = 39) were referred to a moderate or low intensity substance abuse program. The non–enrolled and enrolled groups did not differ in their respective security level designations.

Among those offenders who were referred to SOAW, 15.4% (N = 82) were referred to the program despite not meeting the referral criteria50. The majority of this group met at least some of the criteria (e.g., a rating of high risk on the overall static risk level rating, but fewer than two convictions for a Schedule I offence). Two–way contingency table analyses revealed that 35.8% (N = 29) were scored as high risk to re–offend; of these, 58.6% (N = 17) had prior convictions for two or more Schedule I offences, and more than one–third 34.5% (N = 10) had been convicted of homicide. The remaining 64.2% (N = 52) of offenders who were referred to SOAW despite not meeting all of the criteria were scored as moderate or low risk to re–offend; however, of these offenders, 41.2% (N = 21) had been convicted of at least two Schedule I offences, and 25.5% (N = 13) had been convicted of homicide. As such, analyses suggest that there was a group of offenders deemed suitable for SOAW given their violent histories, despite not meeting all of the specific referral criteria.

Figure 6: Proportion of Offenders Referred to SOAW and Not Enrolled, Women

From a sample of 618 Aboriginal women offender admissions
(Click to enlarge)

[D]

3.2.6 Family Violence Prevention Program Referrals and Enrolments

FINDING 9: There was a pool of candidates who met the referral criteria for a specific family violence prevention program (high or moderate intensity) who were either not referred to the program, or who were referred and did not enrol. Analyses suggest short sentences and high security level designations to be common amongst the non–enrolled groups.

 

Family Violence Prevention Programs

Non–Aboriginal Male Offenders

From a sample of 20,30651 non–Aboriginal male offender admissions, it was determined that 16.6% (N = 3,361) met the referral criteria for either the National Moderate Intensity Family Violence Prevention Program (MIFVPP, N = 2,075) or the HIFVPP (N = 1,286)52. The differences between regions of the proportion of non–Aboriginal male offenders that were referred to a FVPP but not enrolled is seen in Figure 7. Of those offenders who met the referral criteria, 53.5% (N = 1,799) were referred to the program, whereas 46.5% (N = 1,562) were not referred. 18.0% (N = 231) of offenders eligible for HIFVPP were referred to MIFVPP.

A comparison of referrals to non–referrals within the HIFVPP group revealed non–referred offenders as more likely to have shorter sentences (M = 2.83, SD = 2.49 years, versus M = 3.28, SD = 2.45 years, respectively, t(1,053) = –2.94; p < .01). Within the MIFVPP group, no such differences were found. A proportion of HIFVPP and MIFVPP non–referrals were enrolled in SAPs (41.4%, N = 181 and 56.8%, N = 637 respectively), living skills programs (34.3%, N = 151 and 47.7%, N = 535 respectively) and SOPs (22.1%, N = 97 and 12.0%, N = 135 respectively).

Of those offenders referred to either the HIFVPP or the MIFVPP program (N = 1,799), just over one–third enrolled in the program (38.6%, N = 256 and 47.0%, N = 533 respectively). A comparison of the non–enrolments to enrolments revealed that, similar to the eligible non–referrals, the HIFVPP group was more likely to have shorter sentences (M = 3.2, SD = 2.8 years versus M = 3.6, SD = 2.1 years respectively; t(711) = –1.97, p < .05) while no such difference was found for the MIFVPP group. Similar to those who met the referral criteria and were not referred, offenders who were referred to either HIFVPP or MIFVPP and not enrolled were enrolled in SAPs (49.9%, N = 217 and 65.9%, N = 492 respectively), living skills programs (27.8%, N = 121 and 42.7%, N = 319 respectively) and SOPs (16.1%, N = 70 and 15.3%, N = 114 respectively). Those who were not enrolled in a FVPP were more likely to have a maximum security designation than those who enrolled (25.3% vs. 4.0% respectively, x2(2) = 58.98, p < .001 for HIFVPP, and 9.9% vs. 2.1% respectively, x2(2) = 82.21, p < .001 for MIFVPP). Lastly, among those offenders who were referred to a FVPP, only 2.2% (N = 380) were referred to the program despite not meeting the referral criteria.

Figure 7: Proportion of Offenders Referred to FVPPs and Not Enrolled, Non–Aboriginal

Sample of 20,306 non-Aboriginal male offender admissions
(Click to enlarge)

[D]

Aboriginal Male Offenders

From a sample of 4,36353 Aboriginal male offender admissions, it was determined that 32.0% (N = 1,394) met the referral criteria for either the MIFVPP (N = 2,075), HIFVPP (N = 1,286), or the Aboriginal High Intensity Family Violence Program (AHIFVPP)54. For the purposes of the following analyses, the AHIFVPP and HIFVPP have been collapsed into one HIFVPP grouping.

The differences between regions of the proportion of Aboriginal male offenders that were referred to a FVPP but not enrolled is seen in Figure 8. Of those offenders who met the referral criteria, 48.5% (N = 676) were referred to a program, and the remaining 51.5% (N = 718) were not referred. 16.4% (N = 97) of HIFVPP eligible offenders were referred to MIFVPP.

A comparison of referrals to non–referrals within the HIFVPP group revealed no significant differences in sentence length (3.1 versus 3.0 years for HI–VPP respectively, and 2.9 versus 3.2 years for MIFVPP respectively). A large proportion of HIFVPP and MIFVPP non–referrals were enrolled in SAPs (62.8%, N = 155 and 76.2%, N = 359 respectively), living skills programs (24.3%, N = 60 and 41.8%, N = 197 respectively) and SOPs (34.0.1%, N = 84 and 20.4%, N = 96 respectively).

Of those offenders referred to either the HIFVPP or the MIFVPP program (N = 676), just under half enrolled in the program (49.9%, N = 133 and 46.3%, N = 188 respectively). A comparison of non–enrolments to enrolments revealed that, similar to the eligible non–referrals, there were no significant differences in sentence length. Also similar to those who met the referral criteria and were not referred, offenders who were referred to either HIFVPP or MIFVPP and not enrolled were enrolled in SAPs (79.2%, N = 114 and 84.9%, N = 219 respectively), living skills programs (31.3%, N = 45 and 34.1%, N = 88 respectively) and SOPs (16.7%, N = 24 and 15.1%, N = 39 respectively). Those who were not enrolled in a family violence prevention program were more likely to have a maximum security designation than those who enrolled (16.7% vs. 3.7% respectively, x2(2) = 12.63, p < .01 for HIFVPP, and 9.7% vs. 2.9% respectively; x2(2) = 9.59, p < .01 for MIFVPP). Lastly, among those offenders who were referred to a FVPP, only 2.3% (N = 82) were referred to the program despite not meeting the referral criteria.

Figure 8: Proportion of Offenders Referred to FVPPs and Not Enrolled, Aboriginal Male

From a sample of 4,363 Aboriginal male offender admissions
(Click to enlarge)

[D]

3.2.7 Substance Abuse Program Referrals and Enrolments

FINDING 10: Many offenders who met the criteria for a high intensity substance abuse, violence prevention, or sex offender program were referred to a moderate intensity substance abuse program. This discord is reinforced through the most recent changes to substance abuse program referral criteria. There was also a large proportion of offenders who were referred to a substance abuse program who did not enrol.

 

Prior to January 2008, referrals to National Substance Abuse Programs (NSAP) were based on an offenders level of need in the substance abuse domain (as demonstrated by scores on the CLAI/CASA), and an established link between criminal behaviour and substance abuse55. Effective January 14, 2008, static risk was incorporated into substance abuse referrals. However, for the current sample, implementation of substance abuse programming occurred largely prior to the incorporation of risk as a referral criterion. As such, analyses were conducted using the previous criteria of substance abuse need and an established link between substance abuse and crime56.

National Substance Abuse Program – High Intensity

Non–Aboriginal Male Offenders

From a sample of 13,60657 non–Aboriginal male offender admissions, it was determined that 27.6% (N = 3,748) met the referral criteria for the National Substance Abuse Program High Intensity (NSAP–High). The differences between regions of the proportion of non–Aboriginal male offenders that were referred to SAPs but not enrolled is seen in Figure 9. Of those offenders who met the referral criteria, 45.6% (N = 1,709) were referred to the program, whereas 41.5% (N = 1,215) were not referred. A comparison of referrals to non–referrals within this eligible group revealed no differences in average sentence length between the groups; however, the non–referred group was more likely to have either a minimum (19.2% versus 7.7%) or maximum (13.5% versus 7.6%) security level designation when compared to their referred counterparts (x2(2) = 102.0, p < .001). An examination of eligible non–referred offenders revealed that half (50.6%, N = 417) were referred to an education program, and almost one–quarter (22.8%, N = 188) were referred to a living skills program. Lastly, there was a large group of offenders who were not referred to NSAP–High despite meeting the referral criteria, but who were referred to the moderate intensity level of substance abuse programming (58.5%, N = 1,192). Almost one–third of this group also met the criteria for another high intensity program, namely a high intensity VOP or SOP (30.0%, N = 359).

Of those who were referred to NSAP–High (N = 1,709), 35.5% (N = 665) enrolled in the program. A comparison of the non–enrolments to enrolments revealed that, similar to the eligible non–referrals, non–enrolments were more likely to have a maximum security level designation than their enrolled counterparts (11.6% versus 4.4%; x2(2) = 68.7, p < .001) . More than two–thirds of non–enrolled offenders (67.7%, N = 616) were enrolled in an education program, and more than one–quarter (28.5%, N = 259) were enrolled in a living skills program. Of note, more than half (57.7%, N = 910) of offenders who were not enrolled in NSAP–High despite being referred were enrolled in another intensity level of substance abuse programming.

Figure 9: Proportion of Offenders Referred to Substance Abuse Programs and Not Enrolled, Non–Aboriginal Male

From a sample of 13,606 non-Aboriginal male offender admissions
(Click to enlarge)

[D]

Aboriginal Male Offenders

From a sample of 2,89158 Aboriginal male offender admissions, it was determined that 43.1% (N = 1,246) met the referral criteria for the NSAP–High or the Aboriginal Offender Substance Abuse Program (AOSAP)59. The differences between regions of the proportion of non–Aboriginal male offenders that were referred to SAPs but not enrolled is seen in Figure 10. Of those offenders who met the referral criteria, 62.0% (N = 621) were referred to the program, whereas 38.0% (N = 381) were not referred. A comparison of referrals to non–referrals within this eligible group revealed that on average, non–referred offenders were serving shorter sentences than referred offenders (M = 2.82, SD = 1.70 years versus M = 3.42, SD = 2.16 years, respectively; t(562) = –4.31, p < .001), and the non–referred group was more likely to have a minimum (18.5% versus 3.7%) or maximum (18.5% versus 10.5%) security level designation when compared to their referred counterparts (x2(2) = 65.31, p < .001). An examination of eligible non–referred offenders revealed that half (51.2%, N = 125) were referred to an education program, and one–quarter (25.0%, N = 61) were referred to a living skills program. Finally, there was a large group of offenders who were not referred to NSAP–High despite meeting the referral criteria but who were referred to the moderate intensity level of substance abuse programming (60.2%, N = 376). Almost one–third of this group also met the criteria for another high intensity program, namely a high intensity VOP or SOP (58.2%, N = 219).

Of those who were referred to NSAP–High or AOSAP (N = 784), 37.9% enrolled in the program (N = 297). A comparison of the non–enrolments to enrolments revealed that, similar to the eligible non–referrals, non–enrolled offenders were more likely to have a maximum security level designation when compared to their enrolled counterparts (11.8% versus 6.7%, respectively; x2(2) = 14.76, p < .001). Many of the non–enrolled offenders who were referred to, but not enrolled in, either NSAP–High or AOSAP were enrolled in an education program (74.3%, N = 297), a living skills program (35.0%, N = 140), or a FVPP (27.5%, N = 110). Of note, more than half (57.2%, N = 400) of offenders who were not enrolled in NSAP–High or AOSAP despite being referred were enrolled in another intensity level of substance abuse programming.

Figure 10: Proportion of Offenders Referred to Substance Abuse Programs and Not Enrolled, Aboriginal Male

From a sample of 2,891 Aboriginal male offender admissions
(Click to enlarge)

[D]

National Substance Abuse Program Moderate Intensity

Non–Aboriginal Male Offenders

From the same sample of 13,606 non–Aboriginal male offender admissions, it was determined that 15.7% (N = 2,133) met the referral criteria for the National Substance Abuse Program Moderate Intensity (NSAP–Moderate). Of those offenders who met the referral criteria, 71.4% (N = 1,523) were referred to the program. A comparison of referrals to non–referrals within this eligible group revealed no differences in average sentence length between the groups; however, the non–referred group were more likely to have a maximum security level designation when compared to their referred counterparts (11.5% versus 4.9%, respectively; x2(2) = 27.6, p < .001). An examination of eligible non–referred offenders revealed that more than half (52.9%, N = 267) were referred to an education program, and one–quarter (25.4%, N = 128) were referred to a living skills program. Almost one–quarter of offenders (23.0%, N = 116) who were not referred to NSAP–Moderate despite meeting the referral criteria were referred to another intensity level of substance abuse programming.

Of those who were referred to NSAP–Moderate (N = 2,731), 88.1% (N = 2,405) enrolled in the program. A comparison of the non–enrolments to enrolments revealed that, similar to the eligible non–referrals, non–enrolled offenders were more likely to have a maximum security level designation when compared to their enrolled counterparts (10.5% versus 3.3%, respectively; x2(2) = 110.1, p < .001). Many of the non–enrolled offenders who were referred to, but not enrolled in, NSAP–Moderate were enrolled in an education program (67.1%, N = 806) or a living skills program (32.9%, N = 395). Of note, one–third (33.1%, N = 1,202) of offenders who were not enrolled in NSAP–Moderate despite being referred were enrolled in another intensity level of substance abuse programming.

Aboriginal male offenders:
From the sample of 2,891 Aboriginal male offender admissions, it was determined that 20.0% (N = 578) met the referral criteria for the NSAP–Moderate or the AOSAP. Of those offenders who met the referral criteria, 65.7% (N = 380) were referred to the program. A comparison of referrals to non–referrals within this eligible group revealed that on average, non–referred offenders were serving shorter sentences than referred offenders (M = 2.70, SD = 1.75 years versus M = 3.08, SD = 1.57 years, respectively; t(511) = –2.28, p < .05), and the non–referred group were more likely to have a minimum (25.6% versus 16.1%) or maximum (14.1% versus 7.7%) security level designation when compared to their referred counterparts (x2(2) = 11.61, p < .01). An examination of eligible non–referred offenders revealed that more than half (57.9%, N = 77) were referred to an education program, and almost a third (32.3%, N = 43) were referred to a living skills program. One–third of offenders (33.1%, N = 44) who were not referred to NSAP–Moderate or AOSAP despite meeting the referral criteria were referred to another intensity level of substance abuse programming.

Of those who were referred to NSAP–Moderate or AOSAP (N = 987), 38.5% (N = 380) enrolled in the program. A comparison of the non–enrolments to enrolments revealed that, similar to the eligible non–referrals, non–enrolled offenders were more likely to have a maximum security level designation when compared to their enrolled counterparts (16.4% versus 7.3%, respectively; x2(2) = 20.82, p < .001). Many of the non–enrolled offenders who were referred to, but not enrolled in, NSAP–Moderate or AOSAP were enrolled in an education program (66.4%, N = 269) or a living skills program (33.6%, N = 136). Of note, two in five (40.7%, N = 405) offenders who were not enrolled in NSAP–Moderate or AOSAP despite being referred were enrolled in another intensity level of substance abuse programming.

National Substance Abuse Program Low Intensity

Non–Aboriginal Male Offenders

From the same sample of 13,606 non–Aboriginal male offender admissions, it was determined that 27.7% (N = 3,764) met the referral criteria for the National Substance Abuse Program Low Intensity (NSAP–Low). Of those offenders who met the referral criteria, 8.4% (N = 315) were referred to the program. A comparison of referrals to non–referrals within this eligible group revealed that non–referred offenders were serving longer sentences than their referred counterparts (M = 3.2, SD = 2.0 years versus M = 2.7, SD = 0.9 years; t(734) = 7.24, p < .001), and the non–referred group was more likely to have either a medium (59.8% versus 45.7%) or maximum security level designation (7.9% versus 1.3%) when compared to their referred counterparts (x2(2) = 61.73, p < .001). An examination of eligible non–referred offenders revealed that many (63.2%, N = 1,680) were referred to an education program, and one–third (32.7%, N = 869) were referred to a living skills program. Fewer offenders who were not referred to NSAP–Low despite meeting the referral criteria were referred to another intensity level of substance abuse programming (6.4%, N = 170).

Of those who were referred to NSAP–Low (N = 436), 36.9% (N = 161) enrolled in the program. A comparison of the non–enrolments to enrolments revealed that non–enrolled offenders did not differ from enrolled offenders with regard to security level designations (x2(2) = 5.04, p = .08). However, similar to the eligible non–referrals, many offenders who were not enrolled in NSAP–Low despite being referred were enrolled in an education program (66.2%, N = 180) or a living skills program (29.0%, N = 79). Of note, many offenders (62.0%, N = 272) who were not enrolled in NSAP–Low despite being referred were enrolled in another intensity level of substance abuse programming.

Aboriginal Male Offenders

From the sample of 2,891 Aboriginal male offender admissions, it was determined that 25.7% (N = 743) met the referral criteria for the NSAP–Low. Of those offenders who met the referral criteria, 7.9% (N = 59) were referred to the program. A comparison of referrals to non–referrals within this eligible group revealed that on average, non–referred offenders were serving longer sentences than referred offenders (M = 3.00, SD = 1.79 years versus M = 2.64, SD = 0.94 years, respectively; t(127) = 2.40, p < .05), and the non–referred group were more likely to have a medium (61.2% versus 55.9%) or maximum (8.3% versus 3.4%) security level designation when compared to their referred counterparts (x2(2) = 13.12, p < .01). An examination of eligible non–referred offenders revealed that many (68.8%, N = 291) were referred to an education program, and over a third (34.5%, N = 146) were referred to a living skills program. A lower proportion of offenders (16.1%, N = 68) who were not referred to NSAP–Low despite meeting the referral criteria was referred to another intensity level of substance abuse programming.

Of those who were referred to NSAP–Low (N = 71), 40.9% (N = 29) enrolled in the program. A comparison of the non–enrolments to enrolments revealed that groups did not differ with regard to security level designations (x2(2) = 2.49, p = .29). Many of the offenders who were referred to, but not enrolled in, NSAP–Low were enrolled in an education program (72.1%, N = 31) or a living skills program (55.8%, N = 24). More than half (59.7%, N = 43) of offenders who were not enrolled in NSAP–Low despite being referred were enrolled in another intensity level of substance abuse programming.

Women Offender Substance Abuse Program

Intensive Therapeutic Treatment Module

From a sample of 645 women offender admissions, it was determined that 67.3% (N = 434) met the referral criteria for the Intensive Therapeutic Treatment module of the WOSAP60. The differences between regions of the proportion of women offenders that were referred to a SAP but not enrolled is seen in Figure 11. Of those offenders who met the referral criteria, 79.7% (N = 346) were referred to the program, whereas 21.3% (N = 88) were not referred. A comparison of referrals to non–referrals within this eligible group revealed no significant differences in sentence length between groups (t(442) = –1.25, p = .21). An examination of eligible non–referred offenders revealed that many were referred to an education program (70.4%, N = 69) or a living skills program (21.4%, N = 21). Some women offenders who were not referred to the Intensive module despite meeting the referral criteria were referred to another intensity of substance abuse programming (i.e., Module 1 or 3 of WOSAP; 11.2%, N = 11).

Of those who were referred to the Intensive Therapeutic Treatment module of WOSAP (N = 474), 90.1% enrolled in the program (N = 371). A comparison of the non–enrolments to enrolments revealed no significant differences in sentence length (t(492) = –1.42, p = .16). However, the non–enrolled group was more likely to have a maximum security level designation when compared to their enrolled counterparts (11.8% vs. 4.1% respectively, x2(2) = 14.24, p < .001). Many non–enrolled offenders were enrolled in an education program (77.9%, N = 81) or a living skills program (39.4%, N = 41). More than one in five women (21.1%, N = 104) who did not enrol in the Intensive module of WOSAP despite being referred, did enrol in another substance abuse program (i.e., Module 1 or 3 of WOSAP).

Figure 11: Proportion of Offenders Referred to SAPs and Not Enrolled, Women

From a sample of 645 women offender admissions, it was determined that 67.3% (N=434) met the referral criteria
(Click to enlarge)

[D]

Current Substance Abuse Program Referral Criteria

As indicated earlier, prior to January 2008, referrals to NSAP were based on an offenders level of need in the substance abuse domain (as demonstrated by scores on the CLAI/CASA), and an established link between criminal behaviour and substance abuse61. Effective January 14, 2008, static risk was incorporated into substance abuse referrals. As such, the change in referral criteria shifts a proportion of high risk, high need offenders down to requiring moderate intensity level substance–abuse programming. Table 16 demonstrates the proportion of offenders who would shift from a high– to a moderate–intensity substance abuse referral had the criteria been applied to a sample of federal admissions between 1997 and 2007:

Table 16: Current NSAP Referral Criteria (CSC, Reintegration Programs Branch, 2008)62
Assessment Need (CLA/CASA) Level of Risk (SIR or Static Risk Rating) Appropriate Intensity of NSAP
  High 29.5 % (N = 8 928)  High  NSAP–High
4.9 % (N = 1 513)
 High (if offender has been referred to  another high intensity program for a higher  priority need area). Moderate or Low
 NSAP-Moderate 55.1% (N = 7,415)
  Moderate 30.7%    (N = 9,300)  High or Moderate or Low
     Low 39.9% (N = 12 094)  High or Moderate or Low  NSAP–Low 39.9% (N = 12,094)

3.2.8 Sex Offender Program Referrals and Enrolments

FINDING 11: Most non–Aboriginal offenders and half of Aboriginal offenders who met the referral criteria for a specific sex offender program (high, moderate, or low intensity) were referred to the program, and the majority of those referred also enrolled.

 

Sex Offender Programs

Non-Aboriginal Male Offenders

From a sample of 37,57163 non–Aboriginal male offender admissions, it was determined that 15.3% (N = 5,746) met the referral criteria for a High, Moderate or Low Intensity Sex Offender Program (HI–SOP, MI–SOP, or LI–SOP)64. The differences between regions of the proportion of non–Aboriginal male offenders that were referred to a SOP but not enrolled is seen in Figure 12. Of those offenders who met the referral criteria, 61.8% (N = 3,548) were referred to a program, and the remaining 38.2% (N = 2,198) were not referred.

A comparison of non–referrals to referrals within the HI–SOP and MI–SOP groups that met the referral criteria revealed that non–referred offenders were more likely to have shorter sentences (M = 3.3, SD = 2.6 years, versus M = 4.2, SD = 3.5 years, respectively, t(2,200) = –6.73; p < .001 for HI–SOP, and M = 3.3, SD = 2.4 years, versus M = 3.9, SD = 2.2 years, respectively; t(1,191) = –3.08; p < .01 for MI–SOP). Within the LI–SOP group, no such differences were found. A proportion of HI–SOP, MI–SOP, and LI–SOP non–referrals were enrolled in substance abuse programs (56.1%, N = 841, 42.7%, N = 228 and 42.7%, N = 228 respectively) and living skills programs (53.3%, N = 798, 43.1%, N = 230 and 26.5%, N = 44 respectively). Less than 3.0% of non–referrals within all three groups were enrolled in a VOP.

Of those offenders referred to either the HI–SOP, MI–SOP, or the LI–SOP, just over 70% enrolled in a program (N = 2,513). A comparison of the non–enrolments to enrolments revealed no differences in sentence lengths for all SOP groups. Similar to those that met the referral criteria and were not referred, many offenders who were referred to either HI–SOP, MI–SOP OR LI–SOP and did not enrol were enrolled in SAPs (44.3%, N = 173, 41.0%, N = 268 and 38.3%, N = 54 respectively) and living skills programs (38.9%, N = 152, 40.4%, N = 264 and 37.6%, N = 53 respectively). Less than 3.0% of non–referrals within all three groups were enrolled in a violence prevention program. Finally, within those offenders who were referred to a SOP, less than 5% (N = 412) were referred to a SOP despite not meeting the referral criteria.

Figure 12: Proportion of Offenders Referred to SOPs and Not Enrolled, Non–Aboriginal Male

From a sample of 37,571 non-Aboriginal male offender admissions, it was determined that 15.3% (N=5,746) met the referral criteria for a High, Moderate or Low Intensity Sex Offender Program (HI-SOP, MI-SOP, or LI-SOP).
(Click to enlarge)

[D]

 

Aboriginal Male Offenders

From a sample of 8,18365 Aboriginal male offender admissions, 24.9% (N = 2,036) met the referral criteria for HI–SOP, MI–SOP, or LI–SOP66. The differences between regions of the proportion of Aboriginal male offenders that were referred to a SOP but not enrolled is seen in Figure 13. Of those offenders who met the referral criteria, 49.7% (N = 1,011) were referred to a program, and the remaining 50.3% (N = 1,025) were not referred.

A comparison of non–referrals to referrals within the HI–SOP group who met the referral criteria revealed that non–referred offenders were more likely to have shorter sentences (M = 3.2, SD = 2.2 years, versus M = 4.1, SD = 3.1 years, respectively, t(1,052) = –5.24; p < .001). Within the MI–SOP and LI–SOP groups, no such differences were found. A large proportion of HI–SOP and MI–SOP non–referrals was enrolled in substance abuse programs (73.5%, N = 551, and 60.0%, N = 135 respectively), and smaller proportions in these two groups were enrolled in living skills programs (46.4%, N = 348, and 31.1%, N = 70 respectively). Less than 3.0% of non–referrals within all three groups were enrolled in a violence prevention program.

Of those offenders referred to the HI–SOP, MI–SOP, or the LI–SOP, over two–thirds enrolled in a program (68.6%, N = 174). A comparison of the non–enrolments to enrolments revealed no differences in sentence lengths for the HI–SOP and MI–SOP groups, whereas the LI–SOP groups were serving significantly longer sentences (M = 3.4, SD = 1.5 years versus M = 3.0, SD = 0.9 years respectively; t(167) = 2.16, p < .05). Similar to those who met the referral criteria and were not referred, many offenders who were referred to either HI–SOP, MI–SOP, or LI–SOP and did not enrol were enrolled in SAPs (70.5%, N = 79, 83.6%, N = 153 and 70.6%, N = 36 respectively), and living skills programs (48.2%, N = 54, 42.1%, N = 77 and 45.1%, N = 23 respectively). Less than 3.0% of non–enrolments within all three groups were enrolled in a violence prevention program. Finally, within those offenders who were referred to a SOP, less than 2% (N = 86) were referred to a SOP despite not meeting the referral criteria.

Figure 13: Proportion of Offenders Referred to Sex Offender Programs and Not Enrolled, Aboriginal Male

From a sample of 8,183 Aboriginal male offender admissions, 24.9% (N = 2,036) met the referral criteria for HI-SOP, MI-SOP, or LI-SOP. The differences between regions of the proportion of Aboriginal male offenders that were referred to a SOP but not enrolled.
(Click to enlarge)

[D]

3.2.9 Collection and Entering of Information Into OMS

FINDING 12: Institutional and parole office staff suggested that, in general, information is efficiently collected and entered into the Offender Management System during intake, with high efficiency especially pronounced in the Atlantic Region and low efficiency especially pronounced in maximum security institutions.

 

To gauge efficiency, staff members were asked to rate on a scale from one (1) to six (6), with one representing not at all efficient and six representing very efficient, the extent to which information is collected and entered efficiently in order to complete the OIA and Supplementary Intake Assessments. Of those who responded to the question (N = 202), the majority indicated that information collection and recording was done efficiently/very efficiently (36.4%, N = 88) or at least somewhat efficiently (32.6%, N = 79). Only 10.7% (N = 26) indicated that data collection and recording was somewhat inefficient, and 3.7% (N = 9) thought the process was not at all or not efficient (M = 4.29, SD = 1.00).

ANOVAs conducted to determine if any differences existed in mean ratings of efficiency among regions and security levels found significant omnibus differences for both groups (F(4,196) = 3.19, p < .05 and F(5,192) = 5.60, p < .001, respectively). Post–hoc pair–wise comparisons found that, after a Bonferroni correction was applied, the Atlantic Region rated OIA information collection and recording as being significantly more efficient than did either the Prairie (M = 4.62 vs. M = 4.00) and Ontario Regions (M = 4.62 vs. M = 4.03). Post–hoc comparisons of security level data found that those staff in maximum security institutions (M = 3.35) rated OIA information collection and recording as being significantly less efficient than did staff from minimum (M = 4.45), medium (M = 4.17), multi–level (M = 4.81) institutions, and parole office staff (M = 4.39).

For the most part, staff indicated that they could offer recommendations that would aid in increasing the efficiency of the OIA process. Among the most common recommendations was the need for further training and certification of intake assessment staff, greater resources dedicated to the OIA, and a reduction in the use of self–report data in the assessment process.

To garner recommendations for improvement, staff were asked if there was anything that they would recommend that would increase the efficiency of the OIA, Supplementary Intake Assessments, and Correctional Plan development. Of those who answered definitively (yes or no) to the question (N = 200), most (78.0%, N = 156) believed that they could offer recommendations for change (see Table 17 for list of most common recommendations). Eight respondents indicated that they could not offer recommendations but went on to provide comments on OIA efficiency. In total, 164 respondents provided 39 distinct recommendations for change, the most common of which are summarized in Appendix I. A comprehensive list of all recommendations offered can be found in Appendix I.

Table 17: Most Common Recommendations to Improve Efficiency of OIA Process
 Recommendations  Staff Respondents
 More training/certification for intake staff  24 (14.6%)
 Increase the amount of resources  20 (12.2 %)
 Reduce/eliminate the use of self–report data  19 (11.6 %)
 Increase time allotted to complete the OIA  12 (7.3 %)
 Decrease the time taken to obtain community reports  10 (6.1%)
 More teamwork/cooperation throughout the intake  process  10 (6.1%)
 Decrease the number of assessments/instruments  9 (5.5%)

Note: The recommendations listed make up 63.4% of the recommendations noted by staff. Several staff (7.9%, N = 13) made recommendations that did not reflect the topic area in question.

FINDING 13: The majority of staff feel that there is room for improvement when it comes to the OIA process as it related to correctional programs. Increasing intake staff training and certification, and including more CPOs in the intake process were among the most common recommendations for change.

 

Staff members were asked whether there was anything that should be done to improve the OIA process with respect to correctional programs and offender correctional program planning. Of the 199 respondents who answered definitively (yes or no), 157 (78.9%) indicated that there were areas where improvements could be made, and 162 (81.4%) went on to make recommendations for improvement. The most common recommendations included the need for more training and certification of intake staff (11.1%; N = 18), increasing the number of CPOs working at intake (11.1%; N = 18), offering programs for offenders during the OIA (8.6%; N = 14), and increasing the amount of resources that go toward the OIA (8.6%; N = 14). Individual recommendations endorsed by more than 5% of respondents are listed in Table 18, with a complete list of recommendations listed by staff members provided in Appendix I, Table I14.

 

Table 18: Recommendations for Improving the OIA Most Frequently Suggested by Staff Members
 Recommendations  Staff Respondents N (%)
 More training/certification for intake staff  18 (11.1%)
 Have CPOs working at intake  18 (11.1%)
 Offer programs to offenders during OIA  14 (8.6%)
 More resources needed  14 (8.6%)
 Need computer application with program availability  database  12 (7.4 %)
 Shorter assessment periods for short sentences  11 (6.8 %)
 More psychological assessments/info  11 (6.8 %)
 Increase communication between intake staff  10 (6.2 %)
 Have more experienced staff administer the OIA  10 (6.2 %)
 More one–on–one interviews  9 (5.6 %)

Note: The recommendations listed make up 47.6% of those mentioned by staff members.

 

3.2.10 Level of Information Provided to Staff

FINDING 14: Overall, staff members indicated that they have all of the information the need in order to do their job effectively. In addition, staff members indicated that they receive offender information in a timely manner; however this is less the case for staff in maximum security institutions. Police reports and court documents were cited as pieces of information that were both critical in allowing program staff members to do their jobs and as not arriving in a timely manner.

 

Staff members were asked to indicate if they feel that, in their current positions, they have all the offender information they require to do their job, and whether or not they receive offender information in a timely manner.

For those who answered definitively (yes or no; N = 231), the majority of staff respondents (72.7%, N = 168) stated that they have all of the offender information that they require in order to do their job, and that they receive offender information in a timely manner (73.2%, N = 169). Two–way contingency table analyses examining differences in having the required information between regions or security levels found no significant differences. Similarly, two–way contingency table analyses examining differences in timely access to offender information between regions or security levels found that, while regions did not differ in their responses regarding timely access to offender information (x2(4) = 4.30, p = .37), there was a significant omnibus difference between security levels (x2(5) = 16.97, p = .005). Examination of the standardized residuals revealed that respondents from maximum security institutions were more likely to state they did not have timely access to offender information than would be expected by chance (SR = 3.3), and staff from Parole Offices were more likely to state that they do have timely access to offender information than would be expected by chance (SR = 2.0).

For those that answered definitively (yes or no; N = 231), the majority of staff respondents (72.7%, N = 168) believed that they have all of the offender information that they require in order to do their job, and that they receive offender information in a timely manner (73.2%, N = 169). Two–way contingency table analyses examining differences in having the required information between regions or security levels found no significant differences. Similarly, two–way contingency table analyses examining differences in timely access to offender information between regions or security levels found that, while regions did not differ in their responses regarding timely access to offender information (x2(4) = 4.30, p = .37), there was a significant omnibus difference between security levels (x2(5) = 16.97, p = .005). Examination of the standardized residuals revealed that respondents from maximum security institutions were more likely to state they did not have timely access to offender information than would be expected by chance (SR = 3.3), and staff from Parole Offices were more likely to state that they do have timely access to offender information than would be expected by chance (SR = 2.0). In the instances where respondents did not believe that they had the information they required to do their job (27.3%, N = 63) or who believed that they did not receive offender information in a timely manner (26.8%, N = 62), staff were asked to detail the type of information that was required or untimely. Most often, staff stated that they lacked more detailed offender information (such as psychological reports and historical information; 31.7%, N = 20) as well as police reports and court documents (20.6%, N = 13). Police reports and court documents were also noted as being the most common pieces of information that were not provided in a timely manner (32.3%, N = 20), followed by documentation on supplementary intake assessments (16.1%, N = 10). For a comprehensive list of all information detailed as required or untimely by staff, see Appendix I, Tables 1–19.

FINDING 15: The majority of staff report that correctional programs are receiving adequate administrative and management–level support, but feel that support from NHQ is not as adequate. Suggestions to increase support for correctional programs at all levels include increasing communication between administrative, managerial and frontline program staff, emphasizing the importance of correctional programs among all staff members, increasing human and financial resources, and providing updated policy and guideline information.

 

The majority of staff interviewed indicated that correctional programs are receiving adequate support from the administrative level at their site (67.1%, N = 151); more than half indicated adequate support from the management level at their site (56.1%, N = 128), whereas opinions were split evenly as to whether support was adequate at the regional management level (50.0%, N = 101). Fewer staff respondents indicated that correctional programs receive adequate support from NHQ (32.7%, N = 18).

Two–way contingency table analyses revealed no significant differences between regions regarding whether programs receive adequate support at the administrative level, however, respondents from both maximum security institutions (SR = 2.4) and regional offices (SR = 2.9) were significantly more likely to indicate that correctional programs do not receive adequate administrative support (x 2(5) = 19.86, p = .001). Respondents from regional offices were also more likely than would be expected by chance to indicate that correctional programs do not receive adequate support at the management level (SR = 4.4, x2(5) = 25.94, p < .001). There were insufficient data to determine whether any differences exist between regions or security levels regarding support at the NHQ level.

From the 69 respondents who suggested ways in which to increase support for correctional programs at the administrative level, the most common suggestions included increasing funding / human resources (34.8%, N = 24), increasing the number of CPOs / program facilitators (11.6%, N = 8), improving communication between administration and frontline staff (7.2%, N = 5), and offering more programs to offenders (7.2%, N = 5). A comprehensive list of suggestions to increase administrative support for correctional programs is provided in Appendix I, Table I15.

From the 92 respondents who suggested ways in which to increase support for correctional programs at the site or regional management level, the most frequently mentioned suggestions included increasing funding / human resources / direct resources (38.0%, N = 35), improving communication between management and frontline staff (21.7%, N = 20), and increasing staff awareness / emphasizing the importance of correctional programs among staff (16.3%, N = 15). A comprehensive list of suggestions to increase management level support for correctional programs is provided in Appendix I, Table I16.

From the 26 respondents who suggested ways in which to increase support for correctional programs at the NHQ level, the most common suggestions included improving communication between NHQ and frontline / regional staff (23.1%, N = 6), increasing funding / human resources (23.1%, N = 6), increasing NHQ staff members understanding of site operations (19.2%, N = 5), providing research/documentation on correctional programs to program managers (11.5%, N = 3), and updating information and policies in a timely manner / providing guidelines (11.5%, N = 3). A comprehensive list of suggestions to increase NHQ level support for correctional programs is provided in Appendix I, Table I17.

RECOMMENDATION 1: Correctional Service Canada should further support the implementation of correctional programs by ensuring that: (a) information required for the program referral process is obtained from police and courts at an offenders admission, and entered into the Offender Management System in a timely and efficient manner; (b) the intake assessment process is supported by dedicated staff members; (c) offender program assignments are based on sound referral criteria; (d) referrals are prioritized to target multiple risk factors; (e) programs are effectively offered within a time period limited by sentence length and in a restricted maximum security institutional setting; and (f) offenders are encouraged and motivated to engage in their correctional plans.

3.3 Objective 3: Effectiveness:     The extent to which a policy, program, or initiative is meeting its planned results.


SUMMARY FINDING 4: Federal offenders who participated in CSC correctional programs that targeted an identified treatment need exhibited treatment–related change, and were more likely to be granted a discretionary release. In general, correctional program participation was not, however, associated with reductions in major institutional incidents. For community outcomes, program participation was associated with reductions in readmissions, including technical revocations and re-offending (with a non-violent, a violent and/or a sex offence in either federal or provincial jurisdictions). However these findings vary by program, intensity level, and demographic of the federal offender population. Community correctional outcomes were found to be more positive when there was a match between assessed program referral criteria and program intensity level. Conversely, positive effects were diminished when offenders participated in programs that did not correspond to their assessed need. Lastly, positive correctional outcomes were generally more pronounced for offenders who remained enrolled until the end of the program.

FINDING 16: Male offenders who participated in CSC correctional programs that targeted an identified violence prevention treatment need exhibited treatment-related change, were more likely to be granted a discretionary release, and to have positive community correctional outcomes than their comparison group.

3.3.1 Violent Offender Program

Violence Prevention Program

Non-Aboriginal Male Offenders

An examination of institutional outcomes for non-Aboriginal male offenders revealed that VPP participants reported changes in attitudes and beliefs that are consistent with program targets, as all self-reported assessment batteries indicated positive pre-/post-program changes at the p < .05 level or better67. Analyses also revealed that exposure to VPP was associated with an increased chance of being granted a discretionary release. Specifically, non-Aboriginal male offenders in the ITT group were 3.1 times more likely to receive a discretionary release relative to the comparison group (p < .01).

Non-Aboriginal male offenders exposed to VPP were also less likely than the comparison group to be readmitted for any reason (including technical revocations), a new offence, or a new violent offence. Cox Regression analyses were conducted to model the hazard rates of recidivating after being released from incarceration. Results showed that offenders in the ITT group were 29% less likely to be readmitted after release (p < .01), including for technical revocation. ITT group members were 41% less likely to return to custody for a new offence (p < .01), and approximately half as likely (52%, p < .01) to be readmitted with a new violent offence.

When examining the change in the average yearly rate of violent institutional incidents from pre-program to post-program, no significant differences were found between any of the treatment groups and the comparison group of offenders who were not exposed to the program; nor were there any differences between the ITT group as a whole and the comparison group (see Table 19).

When examining the change in the average yearly rate of non-violent institutional incidents from pre-program to post-program, no significant differences were found between any of the treatment groups and the comparison group of offenders who were not exposed to the program; nor were there any differences between the ITT group as a whole and the comparison group.

Table 19: Effectiveness of Violence Prevention for Non-Aboriginal Male Offenders
 Effectiveness Indicator   Intent-to-Treat
N = 502
 Comparison
N = 833
 Institutional Outcomes
    Change in yearly rate of violent     institutional incidents (mean)  –0.03 ns  –0.14 **
    Change in yearly rate of non-violent     institutional incidents (mean)  0.25 ns  –0.33 ns
    Likelihood of discretionary release (Odds     Ratio)  3.05***  1.00
 Community Outcomesa
    Likelihood of any readmission  0.71***  1.00
    Likelihood of readmission with new     offence  0.59***  1.00
    Likelihood of readmission with new violent     offence  0.48***  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

Aboriginal Male Offenders

Similar to male non-Aboriginal program participants, Aboriginal male VPP participants reported changes in attitudes and beliefs that were consistent with program targets, as all self-reported assessment batteries indicated positive pre-/post-program changes at the p < .05 level or better. Further, Aboriginal male offenders who were exposed to VPP were 5.6 times more likely to receive a discretionary release than their Aboriginal male comparison group (p < .01; see Table 20). There were no between-group differences in change in rate of violent or non-violent institutional incidents.

For the community outcomes analyses, trends indicated reductions in readmissions, where offenders in the ITT group were 63% less likely to be readmitted for a new violent offence (p < .10). The results for any readmission and readmissions with a new offence were not statistically significant.

Table 20: Effectiveness of Violence Prevention Program for Aboriginal Male Offenders
 Effectiveness Indicator  Intent-to-Treat  
N = 198
 Comparison
N = 221
 Institutional Outcomes
  Change in yearly rate of violent   institutional incidents (mean)  –0.06 ns  –0.01 ns
  Change in yearly rate of non-violent   institutional incidents (mean)  0.09 ns  0.46 ns
  Likelihood of discretionary release (Odds   Ratio)  5.58***  1.00
 Community Outcomes a
  Likelihood of any readmission  0.82 ns  1.00
  Likelihood of readmission with new offence  0.70 ns  1.00
  Likelihood of readmission with new violent   offence  0.37 *  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

Aggressive Behaviour Control

Non-Aboriginal Male Offenders

An examination of institutional and community outcomes for non-Aboriginal male offenders revealed that exposure to ABC was associated with an increased chance of being granted a discretionary release. Specifically, non-Aboriginal male offenders in the ITT group were 2.3 times more likely than the comparison group to receive a discretionary release relative to the comparison group (p < .05). When examining the change in the average yearly rate of violent and non-violent institutional incidents from pre-program to post-program, no significant differences were found between the ITT group and the comparison group.

Non-Aboriginal male offenders exposed to ABC were also less likely than the comparison group to be readmitted with a new offence. Cox Regression analyses were conducted to model the hazard rates of readmission after being released from incarceration. Results show that offenders in the ITT group were 43% less likely to be readmitted after release for conviction for a new offence (p < .05), and 53% less likely to be readmitted for a new violent offence (p < .10).

Although trends indicate positive results, there were no significant between group differences regarding the likelihood of being readmitted for any reason (see Table 21).

Table 21: Effectiveness of Aggressive Behaviour Control Program for Non-Aboriginal Male Offenders
 Effectiveness Indicator  Intent-to-Treat  
N = 242
 Comparison
N = 238
 Institutional Outcomes
  Change in yearly rate of violent   institutional incidents (mean)  –0.05 ns  –0.03 ns
  Change in yearly rate of non-violent   institutional incidents (mean)  –1.33 ns  –1.02 ns
  Likelihood of discretionary release (Odds   Ratio)  2.27**  1.00
 Community Outcomes a
  Likelihood of any readmission  0.89 ns  1.00
  Likelihood of readmission with new offence  0.57**  1.00
  Likelihood of readmission with new violent   offence  0.47 *  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

Aboriginal Male Offenders

When examining the change in the average yearly rate of violent and non-violent institutional incidents from pre-program to post-program, no significant differences were found between the ITT group and the comparison group of offenders who were not exposed to the program.

The community outcomes analyses revealed that those in the Aboriginal ITT group who were exposed to ABC were 36% (p < .05) less likely to be readmitted for any reason, including technical revocation, and 57% less likely to be readmitted for a new violent offence (p < .10) relative to the Aboriginal comparison group (see Table 22). There were no significant differences between groups regarding the likelihood of returning to custody with a new offence.

Table 22: Effectiveness of Aggressive Behaviour Control Program for Aboriginal Male Offenders
 Effectiveness Indicator  Intent-to-Treat  N = 168  Comparison N = 74
 Institutional Outcomes
  Change in yearly rate of violent institutional   incidents (mean)  –0.09 ns  –0.13*
  Change in yearly rate of non-violent   institutional incidents (mean)  –0.49 ns  –0.39 ns
  Likelihood of discretionary release (Odds   Ratio)  2.79 ns  1.00
 Community Outcomes a
  Likelihood of any readmission  0.64**  1.00
  Likelihood of readmission with new offence  0.71 ns  1.00
  Likelihood of readmission with new violent   offence  0.43 *  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

FINDING 17: Male Aboriginal offenders who participated in violence prevention programs designed for those actively following native spirituality demonstrated improved correctional outcomes.

In Search of Your Warrior

Exposure to ISOYW increased Aboriginal offenders’ likelihood of discretionary release, and reduced their risk for readmission upon release into the community (Table 23). Aboriginal male members in the ITT group (N = 431) were 2.4 times more likely to be granted a discretionary release, relative to the comparison group (p < .01).

Aboriginal male offenders who were exposed to ISOYW were less likely to be readmitted (including for technical revocations) than the comparison group. Cox Regression analyses were conducted to model the hazard rates of recidivating after being released from incarceration. Results indicate that Aboriginal male offenders who were exposed to ISOYW were 19% less likely to be readmitted, relative to offenders in the comparison group (p < .10). There were no significant differences between groups regarding the likelihood of returning with a new offence or with a new violent offence.

When examining the change in the average yearly rate of violent institutional incidents from pre-program to post-program, no significant differences were found between any of the treatment groups and the comparison group of offenders who were not exposed to the program; nor were there any significant differences between the ITT group as a whole and the comparison group.

Table 23: Effectiveness of In Search of Your Warrior for Aboriginal Male Offenders
 Effectiveness Indicator  Intent-to-Treat  N = 431  Comparison N = 289
 Institutional Outcomes
  Change in yearly rate of violent   institutional incidents (mean)  0.00 ns  –0.01 ns
  Change in yearly rate of non-violent   institutional incidents (mean)  –0.32 *  –0.10 ns
  Likelihood of discretionary release (Odds   Ratio)  2.37***  1.00
 Community Outcomes a
  Likelihood of any readmission  0.81*  1.00
  Likelihood of readmission with new offence  0.89 ns  1.00
  Likelihood of readmission with new violent   offence  0.80 ns  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

Spirit of a Warrior68

Aboriginal women offenders who participated in SOAW were more likely to receive a discretionary release compared to offenders in a matched comparison group. There were no differences with regard to involvement in institutional incidents, or likelihood of readmission (see Table 24).

Among Aboriginal women offenders referred to SOAW, the ITT group had significantly fewer institutional charges before program delivery, relative to the comparison group. No other differences were detected between groups, and there were no significant changes between pre- and post-program institutional charge rates for either the ITT group or the comparison group.

Participating in SOAW had a positive impact on offenders likelihood of being granted a discretionary release. Offenders who were exposed to the program (N = 66) were 11.5 times more likely to receive a discretionary release when compared to a similar group of offenders who did not participate in the program (p < .01).

Women offenders exposed to SOAW were not more or less likely than the comparison group to be readmitted to custody after release. Cox Regression analyses were conducted to model the hazard rates of recidivating after being released from incarceration. The ITT and Comparison groups did not differ in their likelihood of readmission for any reason (p = .36) or readmission with a new offence in particular (p = .65). There were insufficient data to analyse whether groups differed in their likelihood to be readmitted for a new violent offence.

Table 24: Effectiveness of Spirit of A Warrior for Aboriginal Women Offenders
 Effectiveness Indicator  Intent-to-Treat  N = 71  ComparisonN = 34
 Institutional Outcomes
  Change in yearly rate of violent institutional   incidents (mean)  –0.02 ns  –0.21 ns
  Likelihood of discretionary release (Odds   Ratio)  11.5***  1.00
 Community Outcomes a
  Likelihood of any readmission  1.35 ns  1.00
  Likelihood of readmission with new offence  0.73 ns  1.00
  Likelihood of readmission with new violent   offence  ––  ––

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

FINDING 18: Male offenders who participated in violent offender programs where the intervention intensity level was congruent with their level of risk demonstrated more positive levels of correctional outcomes. Non-Aboriginal male offenders who participated in violent offender programs where there was not concordance between offender risk and program intensity exhibited no impacts on correctional outcomes. The high intensity violent offender program for Aboriginal male offenders (In Search of Your Warrior) may also be appropriate for Aboriginal offenders requiring a moderate intensity violence prevention program.

Violence Prevention Program

Non-Aboriginal Male Offenders

A review of program assignment investigated whether offenders who met the referral criteria for VPP were correctly identified and subsequently assigned to VPP. Results indicate that some offenders who did not meet the referral criteria were nonetheless assigned to VPP69. As a result, it was decided to analyze institutional and community outcomes separately for those offenders who met the VPP referral criteria and for those offenders who did not.

There were no differences in pre- or post-program rates of violent institutional incidents between ITT members and comparison group members who did not meet the referral criteria for VPP. However, significant differences were found amongst community outcomes.

Non-Aboriginal offenders in the ITT group who met the referral criteria were 37% less likely to return to custody than members of the comparison group (i.e., met the referral criteria for VPP and were not referred). More specifically, ITT group members were half as likely (50%) to be readmitted with a new offence and 43% less likely to be readmitted with a violent offence, relative to comparison group members. Lastly, no between-group differences in community correctional outcomes were found amongst those who did not meet the referral criteria yet were enrolled in VPP.

Aboriginal Male Offenders

Groups did not show significant differences in likelihood of overall readmission, readmission for a new offence, or readmission with a new violent offence. There were insufficient data to analyze group differences in likelihood of readmission between groups for those offenders who did not meet the referral criteria for VPP.

Aggressive Behaviour Control Program

Non-Aboriginal Male Offenders

A review of program assignment investigated whether offenders who met the referral criteria for the ABC were correctly identified and subsequently assigned to this program. Results indicated that some offenders who did not meet the referral criteria were nonetheless assigned to ABC70. As a result, it was decided to analyze institutional and community outcomes separately for those offenders who met the ABC referral criteria and for those offenders who did not.

There were no differences in pre- or post-program rates of violent institutional incidents between ITT members and comparison group members who did not meet the referral criteria for ABC. However, significant differences were found amongst community outcomes.

Non-Aboriginal male offenders in the ITT group who met the referral criteria were 39% less likely to be readmitted for any reason than members of the comparison group (i.e., met the referral criteria for ABC and were not referred). More specifically, ITT group members were 64% less likely to return to custody with a new offence, relative to comparison group members. There were no significant differences between groups regarding the likelihood of returning with a new violent offence. Lastly, no between-group differences in community correctional outcomes were found amongst those who did not meet the referral criteria yet were enrolled in ABC.

Aboriginal Male Offenders

Aboriginal male offenders in the ITT group who met the referral criteria were nearly half as likely (49%) to be readmitted for any reason, including technical revocation, than members of the comparison group (i.e., met the referral criteria for ABC and were not referred). More specifically, ITT group members were 83% less likely to return with a new violent offence, relative to comparison group members. There were no significant differences between groups regarding the likelihood of returning with a new offence. Lastly, no between-group differences in community correctional outcomes were found amongst those who did not meet the referral criteria yet were enrolled in ABC.

In Search of Your Warrior

Participating in ISOYW had a positive impact on offenders who met the referral criteria for the program, and also benefited some offenders who did not meet the specific referral criteria but were enrolled in the program nonetheless. Aboriginal male offenders in the ITT group who met the referral criteria for ISOYW (N = 170) were 6.8 times more likely to receive a discretionary release than comparison group members who met the referral criteria and were not enrolled in the program (p < .0001). In addition, those in the ITT group who did not meet the referral criteria (N = 282) were 3.1 times more likely to be granted discretionary release than comparison group members who did not meet the referral criteria. There were no significant differences between groups regarding the likelihood of readmission for any reason, with a new offence, or a new violent offence.

FINDING 19: Non-Aboriginal male offenders who participated in programs that targeted an identified family violence prevention (intimate partner) treatment need exhibited treatment-related change, and were more likely to be granted a discretionary release. Positive between-group differences for community correctional outcomes were detected for both the high and moderate intensity family violence prevention non-Aboriginal program participants, relative to their respective comparison groups. Additional positive between-group differences were found for community outcomes amongst a sample of moderate intensity program participants whose program intensity levels matched their referral criteria outcomes. No significant between-group differences were found for Aboriginal male program participants71 across institutional and community outcomes.

3.3.2 Family Violence Prevention Programs

Non-Aboriginal Male Offenders

An examination of institutional outcomes for non-Aboriginal male offenders revealed that HIFVPP or MIFVPP participants reported changes in attitudes and beliefs that were consistent with program targets, as 31 of the 32 (97%) of the HIFVPP subscales and all of the 32 MIFVPP subscales (100%) used to measure pre-/post-treatment change were statistically significant and in the expected direction at the p < .0001 level72. Results also indicated that exposure to either HI- or MIFVPP was associated with an increased chance of being granted a discretionary release (3.3 times and 1.7 times more likely than their comparison groups, HIFVPP and MIFVPP respectively). There were significant differences found for violent and non-violent institutional charges for the ITT group (see Table 25).

Non-Aboriginal male offenders exposed to HIFVPP or MIFVPP were also less likely than the comparison group to be readmitted for any reason (including technical revocations; see Table 25 and Table 26). HIFVPP participants were 27% less likely to be readmitted for any reason (p < .05) and MIFVPP participants were 25% less likely to be readmitted for any reason (p < .01).

MIFVPP participants were 36% less likely to be readmitted with a new offense (p < .01) and 57% less likely to be readmitted with a new violent offence (p < .05; see Table 26).

Positive between-group differences were enhanced for community outcomes amongst a sample of moderate intensity program participants whose program intensity levels matched their referral criteria outcomes. Specifically, the ITT group was 25% less likely to be readmitted for any reason (p < .01) than the comparison group.

Table 25: Effectiveness of High-Intensity Family Violence Prevention Program for non-Aboriginal Male Offenders
 Effectiveness Indicator  Intent-to-Treat  N = 487  Comparison N = 476
 Institutional Outcomes    
  Change in yearly rate of violent   institutional incidents (mean)  0.19**  0.06 ns
  Change in yearly rate of non-violent   institutional incidents (mean)  0.41**  0.22 ns
  Likelihood of discretionary release (Odds   Ratio)  3.31***  1.00
 Community Outcomes a    
  Likelihood of any readmission  0.73**  1.00
  Likelihood of readmission with new offence  0.79 ns  1.00
  Likelihood of readmission with new violent   offence  1.67 ns  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

 

Table 26: Effectiveness of Moderate-Intensity Family Violence Prevention Program for non-Aboriginal Male Offenders
 Effectiveness Indicator  Intent-to-Treat  N = 758  Comparison N = 911
 Institutional Outcomes    
  Change in yearly rate of violent   institutional incidents (mean)  0.04 ns  –0.00 ns
  Change in yearly rate of non-violent   institutional incidents (mean)  0.25*  0.13 ns
  Likelihood of discretionary release (Odds   Ratio)  1.70***  1.00
 Community Outcomes a    
  Likelihood of any readmission  0.75***  1.00
  Likelihood of readmission with new offence  0.64***  1.00
  Likelihood of readmission with new violent   offence  0.43**  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

Aboriginal Male Offenders

Similar to non-Aboriginal males, HIFVPP and MIFVPP program participants reported changes in attitudes and beliefs that were consistent with program targets, as 30 of the 32 (94%) HIFVPP subscales and 31 of the 32 (97%) MIFVPP subscales were statistically significant and in the expected direction at the p < .05 level or better.

No significant between-group differences for institutional outcomes (charges and discretionary releases) or community outcomes were found for HIFVPP, MIFVPP or the AHIFVPP groups. Data were not sufficient to conduct these analyses for those who participated in FVPP programs at intensity levels that corresponded to their referral criteria outcome.

FINDING 20: There is a high degree of overlap in programming needs between substance abuse and other programming, which is not reflected in program enrolments. However, analyses suggest that, although not specifically targeting other need areas, substance abuse programs impact needs targeted through other correctional interventions. Overall, male offenders who participated in a substance abuse program exhibited treatment-related change, and were, overall, more likely to be granted a discretionary release than their comparison group. Positive community outcomes were revealed for both Aboriginal and non-Aboriginal male offenders who participated in the National Substance Abuse Programs. Aboriginal offender specific programs targeting substance abuse have positive effects on community outcomes. Lastly, positive between-group differences for community correctional outcomes were detected for male program participants who continued with a substance abuse maintenance program once released into the community.

3.3.3 National Substance Abuse Programs: Overlap of Programming Needs

Programming needs as approximated through program referral criteria73 were compared to actual program enrolment to assess the degree to which offenders with multiple programming needs enrolled in multiple programs. Overall, the highest degree of overlap occurred between substance abuse programming needs and other programs, specifically VOPs, FVPPs, and SOPs. Below is a descriptive analysis comparing the overlap in programming needs with the overlap in actual program enrolments. Each analysis was conducted for high and moderate intensity substance abuse programs separately.

Overlap with Substance Abuse Programs High Intensity74

Non–Aboriginal Male Offenders

An analysis of the non-Aboriginal male offender population determined that while 29.7% (N = 2387) of those who met the High Intensity Violence Prevention Program (HI-VPP)75 referral criteria also met the National Substance Abuse Program, High Intensity (NSAP-High) referral criteria, only 3.7% (N = 33) of enrolments into HI-VPP also enrolled in NSAP-High. Similarly, the overlap in programming needs between the High and Moderate Intensity FVPPs and NSAP-High was 35.3% (N = 1132) while for actual enrolments, this overlap was 5.9% (N = 64). Lastly, the programming needs overlap as established by referral criteria was 13.0% (N = 707) between High, Moderate, and Low Intensity Sex Offender Programs and NSAP-High, while for actual enrolments it was 0.7% (N = 21; see Figure 14).

Aboriginal Male Offenders

Similar to analyses for non-Aboriginal male offenders, the analyses of the Aboriginal male offender population revealed 35.2% (N = 1140) having HI-VPP programming needs, 49.6% (N = 660) having FVPP programming needs, and 28.9% (N = 529) having SOP programming needs also had NSAP-High programming needs. Conversely, only 7.7% (N = 75), 14.8% (N = 58) and 4.0% (N = 34) of enrolments in those programs respectively also enrolled in NSAP-High (see Figure 15).

Figure 14: Programming Needs and Enrolments Overlap with High Intensity Substance Abuse Programming: Non-Aboriginal Male Offenders

An analysis of the non-Aboriginal male offender population determined that while 29.7% (N = 2387) of those who met the High Intensity Violence Prevention Program (HI-VPP)75 referral criteria also met the National Substance Abuse Program, High Intensity (NSAP-High) referral criteria, only 3.7% (N = 33) of enrolments into HI-VPP also enrolled in NSAP-High. Similarly, the overlap in programming needs between the High and Moderate Intensity FVPPs and NSAP-High was 35.3% (N = 1132) while for actual enrolments, this overlap was 5.9% (N = 64). Lastly, the programming needs overlap as established by referral criteria was 13.0% (N = 707) between High, Moderate, and Low Intensity Sex Offender Programs and NSAP-High, while for actual enrolments it was 0.7% (N = 21).
(Click to enlarge)

[D]

Figure 15: Programming Needs and Enrolments - Overlap with High Intensity Substance Abuse Programming: Aboriginal Male OffendersSimilar to analyses for non-Aboriginal male offenders, the analyses of the Aboriginal male offender population revealed 35.2% (N = 1140) having HI-VPP programming needs, 49.6% (N = 660) having FVPP programming needs, and 28.9% (N = 529) having SOP programming needs also had NSAP-High programming needs. Conversely, only 7.7% (N = 75), 14.8% (N = 58) and 4.0% (N = 34) of enrolments in those programs respectively also enrolled in NSAP-High.
(Click to enlarge)

[D]

Women Offenders

The analyses of the women offender population revealed that 67.1% (N = 104) having HI-VPP programming needs and 36.4% (N = 12) having NSAP-High programming needs also had NSAP-High programming needs. Conversely, only 30.8% (N = 41) and 7.7% (N = 1) also enrolled in NSAP-High (see Figure 16).

Figure 16: Programming Needs and Enrolments -- Overlap with High Intensity Substance Abuse Programming: Women Offenders

The analyses of the women offender population revealed that 67.1% (N = 104) having HI-VPP programming needs and 36.4% (N = 12) having NSAP-High programming needs also had NSAP-High programming needs. Conversely, only 30.8% (N = 41) and 7.7% (N = 1) also enrolled in NSAP-High.
(Click to enlarge)

[D]

Overlap with Substance Abuse Programs Moderate Intensity74

Similar to the earlier analyses, Table 27 compares the overlap in programming needs between the Moderate Intensity Substance Abuse Program (NSAP-Moderate) and other programs (VOP, FVPP and SAP) with the overlap in actual program enrolments. The comparisons reveal trends similar to those of the NSAP-High, with the exception of the comparison to FVPP, in which case the trend is reversed (see Table 27).

Table 27: Programming Needs and Enrolments – Overlap with High Intensity Substance Abuse Programming
   Referrals also meeting NSAP-Moderate referral criteria % (N)  Enrolments also enrolled in NSAP-Moderate % (N)
 Program
 Non–Aboriginal Men
         HI–VPP  27.7 (2,113)  12.2 (110)
         FVPP  15.9 (511)  22.0 (238)
         SOP  15.3 (832)  3.2 (95)
 
 Aboriginal Men
         HI–VPP  30.8 (998)  15.6 (153)
         FVPP  17.6 (234)  21.1 (87)
         SOP  27.4 (502)  6.7 (58)
 
 Women
         HI–VPP  17.4 (27)  20.3 (27)
         SOP  ––  ––

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

3.3.4 National Substance Abuse Programs: Effectiveness

National Substance Abuse Program High Intensity

Non-Aboriginal Male Offender

An examination of institutional and community outcomes for non-Aboriginal male offenders revealed that, in comparison to a group of offenders who were referred to NSAP-High but were not enrolled, exposure to NSAP-High was associated with an increased chance of discretionary release (4.5 times as likely). There were no significant group differences regarding offenders’ participation in substance-related institutional incidents from pre- to post-program.

Community outcome analyses revealed positive results, with program participants being 45% less likely to return with a new offence (p < .01) and 63% less likely to return with a new violent offence (p < .01; see Table 28).

Table 28: Effectiveness of NSAP–High for Non–Aboriginal Male Offenders
 Effectiveness Indicator  Intent-to-Treat  N = 611  Comparison N = 668
 Institutional Outcomes
  Change in yearly rate of substance-related   institutional incidents (mean)  0.08 ns  0.09 ns
  Change in yearly rate of non–substance   related institutional incidents (mean)  0.94***  0.21 ns
  Likelihood of discretionary release (Odds   Ratio)  4.51***  1.00
 Community Outcomes a
  Likelihood of any readmission  0.85 ns  1.00
  Likelihood of readmission with new offence  0.55***  1.00
  Likelihood of readmission with new violent   offence  0.37***  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

Aboriginal Male Offenders

Results were similar for Aboriginal male offenders exposed to NSAP-High, with ITT group members having an increased chance of discretionary release (7.2 times as likely) compared to those offenders who were referred to NSAP-High but who were not enrolled. There were no group differences regarding pre- to post-program changes in substance-related institutional incidents. Regarding community outcomes, ITT group members showed no significant differences in likelihood of readmission (37% less likely; p < .10; see Table 29).

Table 29: Effectiveness of NSAP-High for Aboriginal Male Offenders
 Effectiveness Indicator  Intent-to-Treat  N = 147  Comparison N = 195
 Institutional Outcomes
  Change in yearly rate of substance-related   institutional incidents (mean)  0.17 ns  –0.15 ns
  Change in yearly rate of non–substance   related institutional incidents (mean)  –0.15 ns  0.38 ns
  Likelihood of discretionary release (Odds   Ratio)  7.15***  1.00
 Community Outcomes a
  Likelihood of any readmission  0.53 *  1.00
  Likelihood of readmission with new offence  0.61 ns  1.00
  Likelihood of readmission with new violent   offence  0.18 ns  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

National Substance Abuse Program Moderate Intensity

Non-Aboriginal Male Offenders

Participation in NSAP-Moderate was associated with an increased chance (1.4 times more likely) of discretionary release for non-Aboriginal male offenders, relative to a matched group of offenders who were referred to the program but were not enrolled. A repeated-measures Analysis of Variance and post-hoc comparison conducted to determine group differences in the change in annual rate of institutional substance-related incidents revealed that ITT group members had a slight but significant increase in the average rate of incidents relative to the comparison group (see Table 30).

Offenders in the ITT group were 18% less likely to be readmitted for any reason (p < .01; including revocations for violation of a condition of release), 26% less likely to return with a new offence (p < .01), and 46% less likely to be readmitted with a new violent offence when compared to the comparison group (p < .05).

Table 30: Effectiveness of NSAP-Moderate for Non-Aboriginal Male Offenders
 Effectiveness Indicator  Intent-to-Treat
 N = 2,428
 Comparison
 N = 1,079
Institutional Outcomes
  Change in yearly rate of substance-related institutional incidents (mean)  0.10**  –0.03 ns
  Change in yearly rate of non-substance related institutional incidents (mean)  0.31***  –0.26 ns
  Likelihood of discretionary release (Odds Ratio)  1.43***  1.00
 Community Outcomes a
  Likelihood of any readmission  0.82***  1.00
  Likelihood of readmission with new offence  0.74***  1.00
  Likelihood of readmission with new violent offence  0.54**  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

 

Aboriginal Male Offenders

Relative to their comparison group, the NSAP-Moderate ITT group did not differ with respect to substance-related institutional incidents or the likelihood of discretionary release (see Table 31). However, an analysis across community outcomes revealed the Aboriginal male ITT group to be 31% less likely to return with a new offence (p < .10). There were no significant differences between groups regarding the likelihood of readmission for any reason or a new violent offence.

Table 31: Effectiveness of NSAP-Moderate for Aboriginal Male Offenders
 Effectiveness Indicator  Intent-to-Treat  N = 722  Comparison N = 362
 Institutional Outcomes
  Change in yearly rate of substance-related institutional incidents (mean)  –0.02 ns  0.07 ns
  Change in yearly rate of non-substance related institutional incidents (mean)  0.16 ns  –0,46*
  Likelihood of discretionary release (Odds Ratio)  1.37 ns  1.00
 Community Outcomes a
  Likelihood of any readmission  0.85 ns  1.00
  Likelihood of readmission with new offence  0.69*  1.00
  Likelihood of readmission with new violent offence  0.59 ns  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

 

Aboriginal Offender Substance Abuse Program

Relative to offenders who were referred to either AOSAP or NSAP-High but did not participate, participation in AOSAP neither increased nor decreased offenders likelihood of discretionary release, nor did it have an effect on the average rates of substance-related institutional incidents before or after program participation. Community outcomes for AOSAP were positive, with participants half as likely to return to custody after release (p < .05; see Table 32).

Table 32: Effectiveness of AOSAP for Aboriginal Male Offenders
 Effectiveness Indicator  Intent-to-Treat  N = 173  Comparison N = 93
 Résultats en établissement
  Change in yearly rate of substance-related institutional incidents (mean)  0.03 ns  0.03 ns
  Change in yearly rate of non-substance related institutional incidents (mean)  0.38 ns  0.48 ns
  Likelihood of discretionary release (Odds Ratio)  2.21 ns  1.00
 Community Outcomes a
  Likelihood of any readmission  0.51**  1.00
  Likelihood of readmission with new offence  0.49 ns  1.00
  Likelihood of readmission with new violent offence 0.38 ns  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

 

National Substance Abuse Maintenance Program

Non-Aboriginal Male Offenders

Compared to non-Aboriginal male offenders who were referred to the National Substance Abuse Maintenance Program and did not participate (N = 1,886), those who were exposed to the program (N = 3,353) were 22% less likely to be readmitted for any reason (including technical revocations; p < .01) and 27% less likely to be readmitted for a new offence (p < .01; see Table 33). Offenders in the ITT group showed similar chances of returning to custody with a new violent offence when compared to the matched comparison group of offenders who were referred to the NSAP Maintenance Program but did not participate (p = .83).

Table 33: Effectiveness of NSAP Maintenance Program for Non–Aboriginal Male Offenders
 Effectiveness Indicator  Intent-to-Treat  N = 3,353  Comparison N = 1,886
 Community Outcomes a
  Likelihood of any readmission  0.78***  1.00
  Likelihood of readmission with new offence  0.73***  1.00
  Likelihood of readmission with new violent offence  0.97 ns  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

 

Aboriginal Male Offenders

Compared to Aboriginal male offenders who were referred to the NSAP Maintenance Program and did not participate (N = 637), those who were exposed to the program (N = 742) were 24% less likely to be readmitted for any reason (including technical revocations; p < .01) and 30% less likely to return with a new offence (p < .01; see Table 34). Offenders in the ITT group did not significantly differ in their chances of returning to custody with a new violent offence when compared to the matched comparison group of offenders who were referred to the NSAP Maintenance Program but did not participate.

Table 34: Effectiveness of NSAP Maintenance Program for Aboriginal Male Offenders
 Effectiveness Indicator  Intent-to-Treat  N = 742  Comparison N = 637
 Community Outcomes a
  Likelihood of any readmission  0.76***  1.00
  Likelihood of readmission with new offence  0.70***  1.00
  Likelihood of readmission with new violent offence  0.78 ns  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

 

FINDING 21: Male offenders who participated in a sex offender program exhibited treatment-related change, and were, overall, more likely to be granted a discretionary release than their comparison group. Aboriginal males were more likely to demonstrate reductions in major institutional charges respective to the comparison group. Positive community outcomes were revealed for both Aboriginal and non-Aboriginal male offenders who participated in sex offender programs.

3.3.5 Sex Offender Programs

Similar to the other program areas, results of the following analyses were disaggregated by program intensity level (high, moderate, and low) and by Aboriginal and non-Aboriginal status. However, given the unique sub-populations within SOPs, the data were further disaggregated into two victim categories: non-child victims; and child victims (i.e., incest offenders and pedophiles).

Institutional outcome analyses for both non-Aboriginal and Aboriginal male offenders revealed significant changes in pre-/post involvement in major and minor institutional incidents for Aboriginal offenders. Specifically, Aboriginal offenders participating in MI-SOP or LI-SOP demonstrated reductions in major institutional charges while the comparison group experienced increased (M = -0.02 versus M = 0.07, p < .10 for MI-SOP and M = -0.03 versus M = 0.2, p < .10 for LI-SOP). For early discretionary release, results were positive for all intensity levels of SOPs, as well as across location of program (national, regional, or local). Results are summarized in Appendix J

Community outcomes were further divided into the area in which the program was delivered (i.e., national programs, regional programs, and local programs). This level of analysis was performed to indicate the treatment effect of each type of program. Not all locals (national, regional, and local) have significance values reported. This may be the result of three factors: (a) the model may not have been a good fit, in which case values are not computed to that level of detail; (b) there were no observable events for either the intensity level, victim type, or Aboriginal status at that locale; or (c) no program existed at for the specific level of breakdown. Given the multitude of SOP results, these results will be reported together with an indication of the largest contributor to the significance level obtained.

High Intensity Sex Offender Program

Non-Aboriginal Male Offenders with Non-child Victims

An examination of community outcomes for non-Aboriginal male offenders with non-child victims revealed that exposure to HI-SOP was associated with reduced levels of readmissions (25%) and readmission with a new sex offence (71%); however, these results were not statistically significant (see Table 35).

Table 35: Effectiveness of High-Intensity Sex Offender Programs for non-Aboriginal Male Offenders: Non-Child Victims
 Effectiveness Indicator  ITT
All Programs
N = 204
 ITT
National Programs
N = ––
 ITT
Regional Programs
N = 24
 ITT
Local Programs
N = 127
 
Comparison
N = 201
 Community Outcomes a
  Likelihood of any readmission  0.75 ns  ––  0.00  0.69 ns  1.00
  Likelihood of readmission with new offence  0.93 ns  ––  ––  1.02 ns  1.00
  Likelihood of readmission with new violent offence  0.59 ns  ––  ––  0.74 ns  1.00
  Likelihood of readmission with new sex offence  0.29 ns  ––  ––  0.51 ns  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

 

Non-Aboriginal Male Offenders with Child Victims

An examination of community outcomes for non-Aboriginal male offenders with non-child victims revealed that exposure to HI-SOP was associated with reduced levels of readmissions (54%) and readmission with a new sex offence (50%); however, these results were not statistically significant (see Table 36).

Table 36: Effectiveness of High-Intensity Sex Offender Programs for non-Aboriginal Male Offenders: Child Victims
Effectiveness Indicator  ITT
All Programs
N = 249
 ITT
National Programs
N =
 ITT
Regional Programs
N = 93
 ITT
Local Programs
N = 155
 
Comparison
N = 165
 Community Outcomes a
  Likelihood of any readmission  0.46 ns    ––  0.47 ns  1.00
  Likelihood of readmission with new offence  0.36 ns    ––  0.33 ns  1.00
  Likelihood of readmission with new violent offence  0.26 ns    ––  0.33 ns  1.00
  Likelihood of readmission with new sex offence  0.50 ns    ––  0.64 ns  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

 

Aboriginal Male Offenders with Non–child Victims

An examination of community outcomes for Aboriginal male offenders with non-child victims revealed that exposure to HI-SOP was associated with reduced levels of any readmission. Specifically, Aboriginal male offenders with non-child victims in the ITT group were 79% less likely to return with a readmission than the comparison group (p < .01). Results also indicated that the ITT group was less likely to be readmitted with a new offence (79%, p < .05), or a violent offence (93%, p < .05) (see Table 37). Most of the significance contribution can be attributed to both regional and local programs.

Table 37: Effectiveness of High-Intensity Sex Offender Programs for Aboriginal Male Offenders: Non-Child Victims
Effectiveness Indicator  ITT
All Programs
N = 126
 ITT
National Programs
N =
 ITT
Regional Programs
N = 74
 ITT
Local Programs
N = 51
 
Comparison
N = 73
 Community Outcomes a
  Likelihood of any readmission  0.21***    0.28*  0.21***  1.00
  Likelihood of readmission with new offence  0.21***    ––  0.17*  1.00
  Likelihood of readmission with new violent offence  0.07**    ––  0.16 ns  1.00
  Likelihood of readmission with new sex offence  ––    ––  ––  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

 

Aboriginal Male Offenders with Child Victims

An examination of community outcomes for Aboriginal male offenders with child victims revealed no significant findings as compared to the comparison group (see Table 38).

Table 38: Effectiveness of High-Intensity Sex Offender Programs for Aboriginal Male Offenders: Child Victims
 Effectiveness Indicator  ITT
All Programs
N = 92
 ITT
National Programs
N =
 ITT
Regional Programs
N = 59
 ITT
Local Programs
N = 33
 
Comparison
N = 40
 Community Outcomes a
  Likelihood of any readmission  ––    ––  ––  1.00
  Likelihood of readmission with new offence  ––    ––  ––  1.00
  Likelihood of readmission with new violent offence  ––    ––  ––  1.00
  Likelihood of readmission with new sex offence  ––    ––  ––  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

 

Moderate Intensity Sex Offender Program

Non Aboriginal Male Offenders with Non–child Victims

An examination of community outcomes for the non-Aboriginal male offenders with non-child victims revealed that exposure to MI-SOP was associated with a decreased likelihood of any readmission (see Table 39). Specifically, non-Aboriginal male offenders with non-child victims in the ITT group were 28% less likely to be readmitted for any reason than the comparison group (p < .10). Results also indicated that the ITT group was less likely to be readmitted with a new sex offence (70%), but this result was not significant. Most of the significance contribution can be attributed to national programs.

Table 39: Effectiveness of High-Intensity Sex Offender Programs for Aboriginal Male Offenders: Child Victims
 Effectiveness Indicator  ITT
All Programs
N = 500
 ITT
National Programs
N = 232
 ITT
Regional Programs
N = 21
 ITT
Local Programs
N = 247
 
Comparison
N = 295
 Community Outcomes a
  Likelihood of any readmission  0.72*  0.55**  1.65 ns  0.83 ns  1.00
  Likelihood of readmission with new offence  0.56 ns  0.44 ns  ––  0.61 ns  1.00
  Likelihood of readmission with new violent offence  0.46 ns  0.13 ns  ––  0.72 ns  1.00
  Likelihood of readmission with new sex offence  0.30 ns  0.00  ––  0.44 ns  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

 

Non-Aboriginal Male Offenders with Child Victims

An examination of community outcomes for the non–Aboriginal male offenders with child victims revealed that exposure to MI-SOP was associated, at the local level, with a decreased likelihood of any readmission and readmission with a new offence (see Table 40). Specifically, at the local level, non–Aboriginal male offenders with child victims in the ITT group were 60% less likely to be readmitted for any reason than the comparison group (p < .05) and 80% less likely to be readmitted for a new offence (p < .05).

Table 40: Effectiveness of Moderate-Intensity Sex Offender Programs for non-Aboriginal Male Offenders: Child Victims
 Effectiveness Indicator  ITT
All Programs
N = 706
 ITT
National Programs
N = 315
 ITT
Regional Programs
N = 21
 ITT
Local Programs
N = 370
 
Comparison
N = 227
 Community Outcomes a
  Likelihood of any readmission  0.73 ns  1.04 ns  ––  0.40**  1.00
  Likelihood of readmission with new offence  0.64 ns  1.67 ns  ––  0.20**  1.00
  Likelihood of readmission with new violent offence  1.49 ns  0.66 ns  ––  ––  1.00
  Likelihood of readmission with new sex offence  0.75 ns  0.19 ns  ––  ––  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

 

Aboriginal Male Offenders with Non-child Victims

An examination of community outcomes for Aboriginal male offenders with non-child victims revealed that exposure to MI-SOP was associated with reduced levels of readmissions (21%) and readmission with a new sex offence (10%); however, these results were not statistically significant (see Table 41).

Table 41: Effectiveness of Moderate-Intensity Sex Offender Programs for Aboriginal Male Offenders: Non-Child Victims
 Effectiveness Indicator  ITT
All Programs
N = 172
 ITT
National Programs
N = 65
 ITT
Regional Programs
N = 3
 ITT
Local Programs
N = 104
 
Comparison
N = 104
 Community Outcomes a
  Likelihood of any readmission  0.79 ns  1.02 ns  ––  0.71 ns  1.00
  Likelihood of readmission with new offence  0.67 ns  0.55 ns  ––  0.75 ns  1.00
  Likelihood of readmission with new violent offence  2.50 ns  0.00 ns  ––  ––  1.00
  Likelihood of readmission with new sex offence  0.90 ns  0.00 ns  ––  ––  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

 

Aboriginal Male Offenders with Child Victims

An examination of community outcomes for Aboriginal male offenders with child victims revealed that exposure to MI-SOP was not associated with any reduction in readmissions, however, the results were not statistically significant (see Table 42).

Table 42: Effectiveness of Moderate-Intensity Sex Offender Programs for Aboriginal Male Offenders: Child Victims
 Effectiveness Indicator  ITT
All Programs
N = 123
 ITT
National Programs
N = 40
 ITT
Regional Programs
N = ––
 ITT
Local Programs
N = 82
 
Comparison
N = 52
 Community Outcomes a
  Likelihood of any readmission  1.29 ns  0.34 ns  ––  2.53 ns  1.00
  Likelihood of readmission with new offence  ––  ––  ––  ––  1.00
  Likelihood of readmission with new violent offence  ––  ––  ––  ––  1.00
  Likelihood of readmission with new sex offence  ––  ––  ––  ––  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

 

Low Intensity Sex Offender Program

Non–Aboriginal Male Offenders with Non-child Victims

An examination of community outcomes for the non-Aboriginal male offenders with non-child victims revealed that exposure to LI-SOP was associated with a decreased likelihood of any readmission with a new offence, and a readmission with a new violent offence (see Table 43). Specifically, non-Aboriginal male offenders with non-child victims in the ITT group were 86% less likely to return with a readmission with new offence (p < .01), and 91% less likely to return with a new violent offence (p < .10). Most of the significance contribution could be attributed to local programs.

Table 43: Effectiveness of Low-Intensity Sex Offender Programs for non-Aboriginal Male Offenders: Non-Child Victims
 Effectiveness Indicator  ITT
All Programs
N = 158
 ITT
National Programs
N = 79
 ITT
Regional Programs
N = 33
 ITT
Local Programs
N = 46
 
Comparison
N = 57
 Community Outcomes a
  Likelihood of any readmission  0.59 ns  0.41 ns  0.16 ns  0.84 ns  1.00
  Likelihood of readmission with new offence  0.14***  0.24 ns  ––  0.10**  1.00
  Likelihood of readmission with new violent offence  0.09*  ––  2.03 ns  0.00  1.00
  Likelihood of readmission with new sex offence  0.10 ns  ––  2.03 ns  0.00  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

 

Non–Aboriginal Male Offenders with Child Victims

An examination of community outcomes for the non–Aboriginal male offenders with child victims revealed that exposure to LI-SOP was associated with a decreased likelihood of any readmission (see Table 44). Specifically, non–Aboriginal male offenders with child victims in the ITT group were 71% less likely to return with a readmission than the comparison group (p < .01). Most of the significance contribution could be attributed to national and local programs.

Table 44: Effectiveness of Low–Intensity Sex Offender Programs for non–Aboriginal Male Offenders: Child Victims
 Effectiveness Indicator  ITT
All Programs
N = 403
 ITT
National Programs
N = 215
 ITT
Regional Programs
N = 128
 ITT
Local Programs
N = 60
 
Comparison
N = 57
 Community Outcomesa
  Likelihood of any readmission  0.29***  0.22***  ––  0.32**  1.00
  Likelihood of readmission with new offence  0.31 ns  0.33 ns  ––  0.48 ns  1.00
  Likelihood of readmission with new violent offence  ––  7.12 ns  ––  1.26 ns  1.00
  Likelihood of readmission with new sex offence  ––  ––  ––  1.26 ns  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

 

Aboriginal Male Offenders with Non–child Victims

An examination of community outcomes for the Aboriginal male offenders with non-child victims revealed that exposure to LI-SOP was associated with a decreased likelihood of any readmission (see table 45). Specifically, Aboriginal male offenders with non-child victims in the ITT group were 55% less likely to return with a readmission than the comparison group (p < .10). Most of the significance contribution could be attributed to national programs.

Table 45: Effectiveness of Low-Intensity Sex Offender Programs for Aboriginal Male Offenders: Non-Child Victims
 Effectiveness Indicator  ITT
All Programs
N = 58
 ITT
National Programs
N = 37
 ITT
Regional Programs
N = 4
 ITT
Local Programs
N = 17
 
Comparison
N = 37
 Community Outcomes a
  Likelihood of any readmission  0.45*  0.36*  ––  0.92  1.00
  Likelihood of readmission with new offence  0.43 ns  0.32 ns  ––  ––  1.00
  Likelihood of readmission with new violent offence  1.45 ns  ––  ––  ––  1.00
  Likelihood of readmission with new sex offence  0.00  ––  ––  ––  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

 

Aboriginal Male Offenders with Child Victims

An examination of community outcomes for the Aboriginal male offenders with child victims revealed that exposure to LI-SOP was associated with a decreased likelihood of any readmission, readmission with a new violent offence and readmission with a new sex offence (see Table 46). Specifically, Aboriginal male offenders with child victims in the ITT group were 65% less likely to return with a readmission than the comparison group (p < .05), 91% less likely to be readmitted with a new violent offence (p < .10), and 91% less likely to be readmitted with a new sex offence (p < .10). Most of the significance contribution could be attributed to national programs.

Table 46: Effectiveness of Low-Intensity Sex Offender Programs for Aboriginal Male Offenders: Child Victims
 Effectiveness Indicator  ITT
All Programs
N = 56
 ITT
National Programs
N = 37
 ITT
Regional Programs
N = 7
 ITT
Local Programs
N = 12
 
Comparison
N = 17
 Community Outcomes a
  Likelihood of any readmission  0.35**  0.11***  0.00 ns  0.59 ns  1.00
  Likelihood of readmission with new offence  0.29 ns  0.05**  ––  0.00 ns  1.00
  Likelihood of readmission with new violent offence  0.09 ns  0.00 ns  ––  ––  1.00
  Likelihood of readmission with new sex offence  0.09*  0.00 ns  ––  ––  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

FINDING 22: Non-Aboriginal and Aboriginal male offenders who participated in a living skills program exhibited treatment-related change, and were more likely to be granted a discretionary release than members of the comparison group. Positive between-group differences for community correctional outcomes were detected for program participants when compared to their respective comparison groups. As living skills program enrollments overlap with other programs, further analyses revealed other programs such as violent offender programs were the main contributors to the positive outcomes77.

3.3.6 Living Skills Programs

Anger and Emotion Management Program

Non-Aboriginal Male Offenders

An examination of institutional outcomes for non-Aboriginal male offenders revealed that non-Aboriginal male A & EM participants reported changes in attitudes and beliefs that are consistent with program targets, as all self-reported assessment batteries indicated positive pre-/post-program changes at the p < .01 level or better78. Analyses also revealed that exposure to A & EM was associated with an increased chance of being granted a discretionary release and a reduced risk of readmission after release to the community (see Table 53). Specifically, non-Aboriginal male offenders in the ITT group were 1.9 times more likely to receive a discretionary release relative to the comparison group (p < .01). There was no significant change in the rate of violent institutional incidents from pre- to post-program.

Non-Aboriginal male offenders exposed to A & EM were also less likely than the comparison group to be readmitted for any reason (including technical revocations), a new offence, or a new violent offence. Cox Regression analyses were conducted to model the hazard rates of readmission after release. Results show that offenders in the ITT group were 28% less likely to be readmitted after release for any reason (p < .01). ITT group members were 31% less likely to return to custody for a new offence (p < .01), and 34% less likely to be readmitted with a new violent offence than members of the comparison group (p < .01).

Table 47: Effectiveness of Anger & Emotion Management Program for Non-Aboriginal Male Offenders
 Effectiveness Indicator  Intent-to-Treat  N = 3,803  Comparison N = 2,392
 Institutional Outcomes
  Change in yearly rate of violent institutional incidents (mean)  0.01 ns  –0.06 ns
  Change in yearly rate of non–violent institutional incidents (mean)  0.41***  –0.10 ns
  Likelihood of early release (Odds Ratio)  1.88***  1.00
 Community Outcomes a
  Likelihood of any readmission  0.72***  1.00
  Likelihood of readmission with new offence  0.69***  1.00
  Likelihood of readmission with new violent offence 0.66***  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

 

Aboriginal Male Offenders

Similar to male non-Aboriginal program participants, Aboriginal male offenders who were exposed to A & EM reported changes in attitudes and beliefs that were consistent with program targets, as 11 of the 12 (92%) A & EM treatment effect subscales were statistically significant and in the expected direction at the p < .01 level. Aboriginal male participants were also more likely to receive a discretionary release than their Aboriginal male comparison group. There was a slight increase in the rate of violent institutional offences and a significant increase in non-violent offences from pre- to post-program (see Table 54).

Aboriginal male offenders in the ITT group were also less likely than the comparison group to be readmitted for any reason (including technical revocation), a new offence, or a new violent offence. Cox Regression analyses were conducted to model the hazard rates of readmission. Results show that offenders in the ITT group were 18% less likely to be readmitted after release for any reason (p < .01). ITT group members were 24% less likely to return to custody for a new offence (p < .01).

Table 48: Effectiveness of Anger and Emotion Management Program for Aboriginal Male Offenders
 Effectiveness Indicator  Intent-to-Treat  N = 931  Comparison N = 687
 Institutional Outcomes
  Change in yearly rate of violent institutional incidents (mean)  0.06**  –0.11**
  Change in yearly rate of non–violent institutional incidents (mean)  0.49***  0.17 ns
  Likelihood of discretionary release (Odds Ratio)  2.31***  1.00
 Community Outcomes a
  Likelihood of any readmission  0.82***  1.00
  Likelihood of readmission with new offence  0.76***  1.00
  Likelihood of readmission with new violent offence 0.70 ns  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

 

Women Offenders

There were no statistically significant institutional or community outcomes for the women offenders who participated in A & EM (see Table 55), however this may be due to lack of a large enough comparison group.

Table 49: Effectiveness of Anger and Emotion Management Program for Women Offenders
 Effectiveness Indicator  Intent-to-Treat
N = 147
 Comparison
N = 51
 Institutional Outcomes
  Change in yearly rate of violent institutional incidents (mean)  –0.04 ns  –0.36**
  Change in yearly rate of non-violent institutional incidents (mean)  –0.24 ns  0.78 ns
  Likelihood of discretionary release (Odds Ratio)  1.28 ns  1.00
 Community Outcomes a
  Likelihood of any readmission  1.28 ns  1.00
  Likelihood of readmission with new offence  1.53 ns  1.00
  Likelihood of readmission with new violent offence 0.41 ns  1.00

 

Reasoning and Rehabilitation Program

Non–Aboriginal Male Offenders

An examination of institutional outcomes for non-Aboriginal male offenders revealed that non-Aboriginal male R & R participants reported changes in attitudes and beliefs that are consistent with program targets, as 8 of the 9 (89%) R & R subscales were statistically significant and in the expected direction at the p < .0179 . An examination of institutional and community outcomes for non-Aboriginal male offenders revealed that exposure to the R & R program was associated with an increased chance of being granted a discretionary release. Specifically, non-Aboriginal male offenders in the ITT group were 1.7 times more likely to receive a discretionary release relative to the comparison group (p < .01). There was a significant increase in the rate of non-violent institutional incidents and a slight increase in the rate of violent institutional incidents from pre- to post-program (see Table 56).

Non-Aboriginal male offenders exposed to R & R were also less likely than the comparison group to be readmitted for any reason (including technical revocations), a new offence, or a new violent offence. Cox Regression analyses were conducted to model the hazard rates of readmission after being released from incarceration. Results show that offenders in the ITT group were 27% less likely to be readmitted after release for any reason (p < .01). In addition, ITT group members were 31% less likely to return to custody for a new offence (p < .01), and 49% less likely to be readmitted with a new violent offence (p < .01; see Table 56).

Table 50: Effectiveness of Reasoning and Rehabilitation Program for Non-Aboriginal Male Offenders
 Effectiveness Indicator  Intent-to-Treat
N = 3,066
 Comparison
N = 1,218
 Institutional Outcomes
  Change in yearly rate of violent institutional incidents (mean)  0.02*  –0.03 ns
  Change in yearly rate of non-violent institutional incidents (mean)  0.56***  0.04 ns
  Likelihood of early release (Odds Ratio)  1.65**  1.00
 Community Outcomes a
  Likelihood of any readmission  0.73***  1.00
  Likelihood of readmission with new offence  0.69***  1.00
  Likelihood of readmission with new violent offence 0.51***  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

 

Aboriginal Male Offenders

Similar to male non-Aboriginal program participants, Aboriginal male offenders who were exposed to R & R reported changes in attitudes and beliefs that are consistent with program targets, as 8 of the 9 (89%) subscales were statistically significant and in the expected direction at the p < .01 level. Further, Aboriginal R & R participants were 1.7 times more likely to receive a discretionary release than their Aboriginal male comparison group. There was also an increase in the rate of non-violent institutional incidents from pre- to post-program, but no significant increase in the rate of violent institutional incidents (see Table 57).

Similar to the non-Aboriginal male program participants, Aboriginal male offenders were less likely to be readmitted for any reason (including technical revocation). Cox Regression analyses were conducted to model the hazard rates of readmission. Results show that offenders in the ITT group were 19% less likely to be readmitted after release for any reason (including for technical revocations; p < 05). ITT group members were neither more nor less likely than comparison group members to return to custody following release for a new offence or a new violent offence (Table 57).

Table 51: Effectiveness of Reasoning and Rehabilitation Program for Aboriginal Male Offenders
 Effectiveness Indicator  Intent-to-Treat
N = 595
 Comparison
N = 267
 Institutional Outcomes
  Change in yearly rate of violent institutional incidents (mean)  0.00 ns  0.11 ns
  Change in yearly rate of non-violent institutional incidents (mean)  0.42**  0.69 ns
  Likelihood of discretionary release (Odds Ratio)  1.65***  1.00
 Community Outcomes a
  Likelihood of any readmission  0.81**  1.00
  Likelihood of readmission with new offence  0.86 ns  1.00
  Likelihood of readmission with new violent offence 0.76 ns  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

 

Women Offenders

Women offenders who participated in R & R had a significant decrease in major charges after program participation. There was a slight decrease in the rate of violent institutional incidents from pre-to post-program. No other statistically significant institutional or community outcomes were revealed, however this may be due to lack of a large enough comparison group (see Table 58).

Table 52: Effectiveness of Reasoning and Rehabilitation Program for Women Offenders
 Effectiveness Indicator  Intent-to-Treat
 N = 66
 Comparison
N = 17
 Institutional Outcomes
  Change in yearly rate of violent institutional incidents (mean)  0.34*  0.43 ns
  Change in yearly rate of non-violent institutional incidents (mean)  –1.22 ns  0.85 ns
  Likelihood of discretionary release (Odds Ratio)  1.42 ns  1.00
 Community Outcomes a
  Likelihood of any readmission  0.67 ns  1.00
  Likelihood of readmission with new offence  0.97 ns  1.00
  Likelihood of readmission with new violent offence 0.00 ns  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

 

FINDING 23: Male and female offenders who participated in a community maintenance program were more likely to have positive community correctional outcomes when compared to their respective comparison groups.

3.3.7 Community Programs

Community Maintenance Program

Non-Aboriginal Male Offenders

An examination of community outcomes for non-Aboriginal male offenders revealed that exposure to the Community Maintenance Program (CMP) was associated with a reduced risk of readmission for any reason, including technical revocations, new offences, or new violent offences after release into the community (see Table 59). Cox Regression analyses were conducted to model the hazard rates of readmission after being released from incarceration. Results show that offenders in the ITT group were 29% less likely than the comparison group to be readmitted after release for any reason, including for technical revocation (p < .01). ITT group members were 40% less likely to return to custody for a new offence (p < .01), and 56% less likely to be readmitted for a new violent offence than members of the comparison group (p < .01).

Table 53: Effectiveness of Community Maintenance Program for Non-Aboriginal Male Offenders
 Effectiveness Indicator  Intent-to-Treat
N = 430
 Comparison
N = 586
 Community Outcomesa
  Likelihood of any readmission  0.71***  1.00
  Likelihood of readmission with new offence  0.60***  1.00
  Likelihood of readmission with new violent offence  0.44***  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

 

Aboriginal Male Offenders

Similar to male non-Aboriginal program participants, Aboriginal male offenders who were exposed to CMP were less likely than the comparison group to be readmitted for any reason (including technical revocations), a new offence, or a new violent offence (see Table 60). Cox Regression analyses were conducted to model the hazard rates of readmission after release. Results show that offenders in the ITT group were 42% less likely than the comparison group to be readmitted for any reason, including for technical revocation (p < .01). ITT group members were 53% less likely to return to custody for a new offence (p < 01), and 59% less likely to be readmitted for a new violent offence than members of the comparison group (p < .10).

Table 54: Effectiveness of Community Maintenance Program for Aboriginal Male Offenders
 Effectiveness Indicator  Intent-to-Treat
N = 102
 Comparison
N = 207
 Community Outcomes a
  Likelihood of any readmission  0.58***  1.00
  Likelihood of readmission with new offence  0.47***  1.00
  Likelihood of readmission with new violent offence  0.41*  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

 

Women Offender Relapse Prevention Program

An examination of community outcomes for women offenders revealed that women offenders exposed to the Women Offender Relapse Prevention Program were less likely than the comparison group to be readmitted for any reason (including technical revocations; see Table 61). Cox Regression analyses were conducted to model the hazard rates of readmission after release. Results show that offenders in the ITT group were 38% less likely than the comparison group to be readmitted for any reason, including for technical revocation. There were no significant differences between groups regarding the likelihood of returning for a new offence or for a new violent offence.

Table 55: Effectiveness of Women Offender Relapse Prevention Program for Women Offenders
 Effectiveness Indicator  Intent-to-Treat
N = 271
 Comparison
N = 115
 Community Outcomesa
  Likelihood of any readmission  0.62***  1.00
  Likelihood of readmission with new offence  0.78 ns  1.00
  Likelihood of readmission with new violent offence  0.50 ns  1.00

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01.

 

RECOMMENDATION 2:

To maximize outcomes attributed to correctional programming, CSC should prioritize programs based on level of risk, offer programs to offenders at intensity levels that are commensurate to sound referral criteria ratings, and develop a strategy to address the violence prevention programming needs of those offenders who also have substance abuse needs.

3.4 Objective 4: Cost Efficiency: The extent to which correctional programs are producing planned outputs in relation to financial resources used.

SUMMARY FINDING 5: Correctional programs are delivered in federal institutions cost-efficiently, as between 70 to 90 cents of every dollar spent on programs is allocated towards successful completions. The community maintenance programs examined were less cost-efficient, as only 30 to 60 cents per dollar were spent on successful completions.

 

Violent Offender Programs

Over 75% of the allocated costs to VOPs in fiscal year 2006/2007 were associated with successful completions. In the case of SOAW, over 90% of the cost were associated with successful completions.

Figure 17: Cost Allocation for Violent Offender Programs

A comparison of completions, drop outs, and administrative procedures for Violent Offender Programs for the 2006/2007 Fiscal Year.  Over 75% of the allocated costs were associated with successful completions. In the case of SOAW, over 90% of the cost were associated with successful completions.
(Click to enlarge)

[D]

Family Violence Prevention Programs

Over 70% of the total costs allocated to the FVPPs were associated with successful completions. The maintenance program had the lowest proportion at just over 70%, while nearly 85% of allocated costs for the HIFVPP were associated with successful completions.

Figure 18: Cost Allocation for Family Violence Programs

A comparison of completions, drop outs, and administrative procedures for Family Violence Prevention Programs for the 2006/2007 Fiscal Year.  Over 70% of the total costs allocated to the FVPPs were associated with successful completions. The maintenance program had the lowest proportion at just over 70%, while nearly 85% of allocated costs for the HIFVPP were associated with successful completions.
(Click to enlarge)

[D]

Living Skills Programs

For both the R & R and the A & EM programs, successful completions made up approximately 80% of the allocated costs for each program, while 19% of the costs were associated with offenders who dropped out of the programs.

Figure 19: Cost Allocation for Living Skills Programs

A comparison of completions, drop outs, and administrative procedures for Living Skills Programs for the 2006/2007 Fiscal Year.  For both the R & R and the A & EM programs, successful completions made up approximately 80% of the allocated costs for each program, while 19% of the costs were associated with offenders who dropped out of the programs.
(Click to enlarge)

[D]

Sex Offender Programs

There was a substantial difference across sex offender programs with respect to the proportion of costs allocated towards successful program completion. While for the MI-SOP and the LI-SOP, successful completions made up approximately 90% of the allocated costs for each program, only 50% of costs for the Maintenance SOP and the Sex Offender Therapy Program for Women were associated with successful completions. However, in the latter two cases, the proportion of drop-outs were low to zero, while the vast majority of non-completions were due to population management (i.e., administrative issues such as the transfer or release of offenders).

Figure 20: Cost Allocation for Sex Offender Programs

A comparison of completions, drop outs, and administrative procedures for Sex Offender Programs for the 2006/2007 Fiscal Year.  There was a substantial difference across sex offender programs with respect to the proportion of costs allocated towards successful program completion. While for the MI-SOP and the LI-SOP, successful completions made up approximately 90% of the allocated costs for each program, only 50% of costs for the Maintenance SOP and the Sex Offender Therapy Program for Women were associated with successful completions. However, in the latter two cases, the proportion of drop-outs were low to zero, while the vast majority of non-completions were due to population management (i.e., administrative issues such as the transfer or release of offenders).
(Click to enlarge)

[D]

Substance Abuse Programs

There was a substantial difference across substance abuse programs with respect to the proportion of costs allocated towards successful program completion. While for four of the programs (NSAP-High, NSAP-Moderate, NSAP-Low and AOSAP), successful completions made up 70% to 85% of the allocated costs for each program, 40-55% of costs for the National Substance Abuse Maintenance Program and WOSAP were associated with successful completions. In each of the latter cases, nearly 30% of the costs were associated with population management (i.e., administrative issues such as the transfer or release of offenders).

Figure 21: Cost Allocation for Substance Abuse Programs

A comparison of completions, drop outs, and administrative procedures for Substance Abuse Programs for the 2006/2007 Fiscal Year.  There was a substantial difference across substance abuse programs with respect to the proportion of costs allocated towards successful program completion. While for four of the programs (NSAP-High, NSAP-Moderate, NSAP-Low and AOSAP), successful completions made up 70% to 85% of the allocated costs for each program, 40-55% of costs for the National Substance Abuse Maintenance Program and WOSAP were associated with successful completions. In each of the latter cases, nearly 30% of the costs were associated with population management (i.e., administrative issues such as the transfer or release of offenders).
(Click to enlarge)

[D]

Community Programs

Of the costs allocated to the Community Maintenance Program, 43% were associated with successful completions, while 47% of the costs were associated with offenders who dropped out of the program.

Figure 22: Cost Allocation for Community Programs

A comparison of completions, drop outs, and administrative procedures for Community Programs  for the 2006/2007 Fiscal Year. Of the costs allocated to the Community Maintenance Program, 43% were associated with successful completions, while 47% of the costs were associated with offenders who dropped out of the program.
(Click to enlarge)

[D]

Women Offender Program

Of the costs allocated to the Women Offender Relapse Program, just over 30% were associated with successful completions, and roughly 55% were associated with offenders who dropped out of the program. However, these results are for FY 06/07, and there was a significant improvement for this year over previous years.

Figure 23: Cost Allocation for Women Offender Programs

A comparison of completions, drop outs, and administrative procedures for the Women Offender Program (Women Offender Relapse) for the 2006/2007 Fiscal Year. Of the costs allocated to the program, just over 30% were associated with successful completions, and roughly 55% were associated with offenders who dropped out of the program. However, these results are for FY 06/07, and there was a significant improvement for this year over previous years.
(Click to enlarge)

[D]

3.5 Objective 5: Cost-Effectiveness: Cost-effectiveness determines the relationship between the amount spent and the results achieved relative to alternative design and delivery approaches.

The cost effectiveness-analyses presented in this section associate correctional program costs to both institutional outcomes (discretionary release) and community outcomes (readmission). Note that the outcomes were converted from likelihood ratios to more tangible measures to facilitate analyses. Specifically, what was reported earlier for correctional program participants as a likelihood of receiving an early (discretionary) release was converted to the number of incarceration days reduced, compared to the non-program participant comparison group. Similarly, the likelihood of remaining in the community without returning (technical revocation or re-offending) was converted to the number of days participants successfully remained in the community.

Next, cost savings were associated to outcomes as a function of avoided incarceration days, and extended community supervision days (up until the end of the federal sentence). Specifically, the ITT and comparison groups were compared, where calculations considered the daily cost to incarcerate an inmate, offset by the extended daily cost of supervising an offender in the community. For example, the cost of a reduction of one incarceration day is offset by the cost of one resulting day spent in community under supervision. Similarly, a cost was attached to community outcomes, where the cost of an additional day of supervision was offset by the avoided cost of a day of incarceration. Finally, cost savings (positive or negative) were compared to the unit costs of delivering the target correctional program, using fiscal 2006/2007 as the base year.

SUMMARY FINDING 6: The vast majority of correctional programs evaluated were cost-effective as, on average, every dollar spent on programming resulted in a return ranging from 1 to 8 dollars. Returns represent cost savings of fewer days of incarceration. This reduction in incarceration costs was, in turn, attributed to earlier releases and extended stays in the community.

FINDING 24: Violent offender programs are cost-effective, as every dollar CSC spends on violent offender programming results in a return 80of 2 dollars and 26 cents.

Violent Offender Programs

Cost effective analyses reveal that every dollar spent on VOP yields $2.26 in cost savings. Approximately 55% of this return is attributed to the savings associated to reduced incarceration time, and 45% to savings attributed to reduced readmission. Note that this ratio varies by program. A break down of cost effectiveness by program is displayed in Table 62.

Table 56: Cost Effectiveness by Violent Offender Program


 Program
 

Savings
 Program Delivery Cost81  
Return on Expenditure
 

Distribution of Return
  Institution Community Unit Cost One Dollar Institutional Community
 All Violent  Offender  Programs82  
$10,856.66
 
$8,898.90
 
$8,755.00
 
$2.26
 
54.95%
 
45.05%
 Violence Prevention Program  
$7,831.04
 
$9,610,82
 
10 749,00 $
 
$1.62
 
44.90%
 
55.10%
 In Search of Your Warrior  
$12,992.45
 
$5,339.34
 
$5,269.00
 
$3.48
 
70.87%
 
29.13%
 Spirit of A Warrior  
$11,568.00
 
$0.00
 
$6,145.00
 
$1.88
 
100.00%
 
0.00%

 

FINDING 25: Family violence prevention programs are cost-effective. On average, every dollar CSC spends on family violence prevention programming results in a return of 1 dollar and 33 cents.

Family Violence Prevention Programs

Cost effectiveness analyses reveal that every dollar spent on FVPPs yields, on average, $1.33 in cost savings. Approximately 84% of this return is attributed to the savings associated with reduced incarceration time, and 16% to savings associated with reduced readmission. Note that this ratio varies by program. A break down of cost effectiveness by program is displayed in Table 63.

Table 57: Cost Effectiveness by Family Violence Prevention Program


 Program
 

Savings
 Program Delivery Cost  
Return on Expenditure
 

Distribution of Return
  Institution Community Unit Cost One Dollar Institutional Community
 All Family  Violence  Prevention  Programs83  
$5,517.38
 
$1,067.88
 
$4,969
 
$1.33
 
83.78%
 
16.22%
 Moderate  Intensity  FVPP  
$4,805.46
 
$711.92
 
$2,391
 
$2.31
 
87.10%
 
12.90%
 High  Intensity  FVPP  
$11,212.74
 
$3,203.64
 
$5,963
 
$2.42
 
77.78%
 
22.22%
 Aboriginal  High  Intensity  FVPP  
––
 
––
 
––
 
––
 
––
 
––

 

FINDING 26: Substance abuse programs are cost-effective, as every dollar spent on substance abuse programming results in an average return of 2 dollars and 69 cents.

Substance Abuse Programs

Cost effectiveness analyses reveal that every dollar spent on SAPs yields, on average, $2.69 in cost savings. Approximately 49% of this return is attributed to the savings associated with reduced incarceration time, and 51% to savings associated with reduced readmission. Note that this ratio varies by program. Notably, the return on expenditure for the NSAP-Moderate is 100% attributed to institutional outcomes (early release). It should be noted that the likelihood for this group to be readmitted is 18% lower then that of the comparison group. However, the offenders who participated in the program were readmitted sooner than the comparison group. A more in-depth analysis reveals that this is the case only for technical revocations and not for new offences or new violent offences. A potential explanation for this is that there is a high proportion of high risk, high need offenders participating in this program (see implementation section). A breakdown of cost-effectiveness by program is displayed in Table 64.

Table 58: Cost Effectiveness by Substance Abuse Program


 Program
 

Savings
 Program Delivery Cost  
Return on Expenditure
 

Distribution of Return
  Institution Community Unit Cost One Dollar Institutional Community
 All Male  Substance  Abuse  Programs84  
$6,585.26
 
$6,763.24
 
$4,969
 
$2.69
 
49.33%
 
50.67%
 NSAP–High  
$9,610.92
 
$4,271.52
 
$6,845
 
$2.03
 
69.23%
 
30.77%
 NSAP– Moderate  
$4,449.50
 
$0.00
 
$1,977
 
$2.25
 
100.00%
 
0.00%
 AOSAP  
$0.00
 
$3,916.00
 
$1,089
 
$3.60
 
0.00%
 
100.00%
 Community  Maintenance  
$0.00
 
$6,763.24
 
$1,602
 
$4.22
 
0.00%
 
100.00%

FINDING 27: Sex offender programs are cost-effective. Every dollar CSC spends on sex offender prevention programming results in, on average, a return of 6 dollars and 59 cents.

Sex Offender Programs

Cost effectiveness analyses reveal that every dollar spent on sex offender programs yields, on average, $6.59 in cost savings. Approximately 60% of this return is attributed to the savings associated with reduced incarceration time, and 40% to savings associated with reduced readmission. Note that this ratio varies by program. A breakdown of cost-effectiveness by program is displayed in Table 65.

Table 59: Cost Effectiveness by Sex Offender Program


 Program
 

Savings
 Program Delivery Cost  
Return on Expenditure
 

Distribution of Return
  Institution Community Unit Cost One Dollar Institutional Community
 All Male Sex  Offender  Programs85  
$19,755.78
 
$12,992.54
 
$4,969
 
$6.59
 
60.33%
 
39.67%
 Low  Intensity Sex  Offender  Program  
––
 
––
 
$745
 
––
 
––
 
––
 Moderate  Intensity Sex  Offender  Program  
$18,865.88
 
$12,458.60
 
$9,189
 
$3.41
 
60.23%
 
39.77%
 High  Intensity Sex  Offender  Program  
$9,432.94
 
$11,568.70
 
––
 
––
 
––
 
––

FINDING 28: Living skills programs are cost-effective. Every dollar CSC spends on living skills programming results in, on average, a return of 3 dollars and 73 cents.

 

Living Skills Programs

Cost effectiveness analyses reveal that every dollar spent on living skills programs yields, on average, $3.73 in cost savings. Approximately 62% of this return is attributed to the savings associated with reduced incarceration time, and 38% to savings associated with reduced readmission that this ratio varies by program. A breakdown of cost-effectiveness by program is displayed in Table 66.

Table 60: Cost Effectiveness by Living Skills Program


 Program
 

Savings
 Program Delivery Cost  
Return on Expenditure
 

Distribution of Return
  Institution Community Unit Cost One Dollar Institutional Community
 All Male  Offender  Living Skills  Programs86  
$11,390.72
 
$7,119.20
 
$4,969
 
$3.73
 
61.54%
 
38.46%
 Anger and  Emotion  Management  
$11,390.72
 
$8,721.02
 
$2,471
 
$8.14
 
56.64%
 
43.36%
 Reasoning  and  Rehabilitation  
$11,390.72
 
$4,627.48
 
$3,968
 
$4.04
 
71.11%
 
28.89%

FINDING 29: Most staff had suggestions for best practices in the implementation of correctional programs, with suggestions centering around the need for more resources for correctional programs, more staff training, and more programming overall.

As a final wrap-up to the interview, staff members were asked whether they had any additional suggestions for best practices in the implementation of correction programs. Of the 206 respondents who answered definitively (yes or no), 160 (77.7%) felt that they could offer suggestions for best practices, and 174 (84.5%) went on to make recommendations for change. The most common recommendations for best practices included the need for more resources in the way of funding, staff, and staff support (20.1%; N = 35), the need for more staff training and mentoring (19.0%; N = 33), and the need for more programming at all security levels (15.5%; N = 27). Individual recommendations endorsed by more than 5% of respondents are listed in Table 67, with a complete list of recommendations listed by staff members provided in Appendix I, Table I18.

Table 61: Summary of Suggestions for Best Practices in the Implementation of Correctional Programs
 Recommendations:  Staff Respondents N (%)
 More resources ($, staff, staff support)  35 (20.1%)
 More staff training/mentoring  33 (19.0%)
 More programs at all security levels  27 (15.5%)
 More flexibility in program delivery  24 (13.8%)
 More communication b/w staff/sites/NPB  22 (12.6%)
 More community support/ programs  20 (11.5%)
 Hire professional/certified staff/degree requirement  12 (6.9%)
 Move to modular program delivery  12 (6.9%)
 More accuracy in program placement/assessment  11 (6.3%)
 More set/continuous program schedules  10 (5.7%)
More 1-on-1 programming/individual inter  10 (5.7%)
 More team–oriented approach  10 (5.7%)
 More flexibility in program referral/placement  10 (5.7%)

Note: The recommendations listed make up 76.1% of those mentioned by staff members.

RECOMMENDATION 3:

Given the positive results of CSC programs in terms of effectiveness, cost-efficiency, and cost-effectiveness, CSC should maximize its return on program expenditures by allocating resources to facilitate the early identification of programming needs, refer more offenders to programs commensurate with their prioritized programming needs, and increase participation rates of offenders who are referred to a program yet do not enroll. Considering the positive outcomes for all offenders participating in community maintenance programs, CSC should also consider strengthening this strategy such that there are more successful completions.

3.6 Objective 6: Continued Relevancy: The extent to which correctional programs continue to realistically addresses an actual need.

SUMMARY FINDING 7: Correctional programs continue to address the risk levels and specific criminogenic needs of the current offender population, and thus continue to be a relevant, effective means of enhancing public safety and reducing re-offending. Nonetheless, CSC is facing challenges regarding the effective and efficient delivery of correctional programs which will address the increased proportion of offenders with education and learning deficits, mental disorders, shorter sentences, and higher security level ratings.

An operational strategy is considered relevant when it remains consistent with government-wide and departmental priorities, and continues to address an actual need. The present evaluation of correctional programs thus considered the changing offender profile and its potential impacts on programming needs. There is an indirect link between changing demographic profiles and correctional programs given that the changing demographics impact offenders risk/need profiles (e.g., younger offenders are generally higher risk than older offenders), which in turn inform program referrals. Demographic characteristics can also have implications for responsivity to treatment. For example, younger offenders with lower education levels might require educational interventions prior to participating in correctional programs, so that they may benefit from the program content. Alternatively, programs may need to consider changes to program delivery methods in order to be appropriately responsive to offender characteristics.

Overall, the present evaluation shows that participation in correctional programs can be associated with positive community correctional outcomes, as evidenced by the results for male offenders (see Table 68)87.

Table 62: Community Outcomes for All Programs: Male Offenders
 Table 62: Community Outcomes for All Programs: Male Offenders
   VOP  FVPP  NSAP  SOP(a)  SOP(b)  LSP
 Community Outcomes a
 Any Return  0.76***  0.76***  0.80***  0.64***  0.60***  0.75***
 Readmission with a new  offence
 0.70***

 0.76***

 0.68***

 0.55***

 0.74 ns

 0.72***
 Readmission with a new  violent offence
 0.52***

 0.72 ns

 0.45***

 0.59*

 0.78 ns

 0.63***
 Readmission with a new sex  offence
 ––

 ––

 ––

 0.59 ns

 0.98 ns

 ––

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01. The above  results are not replicated in other areas of the report, since figures aggregate the male  non-Aboriginal and Aboriginal populations, as well as multiple intensity levels within  program.

The following is a brief description of trends in the demographic profile of federal offenders and possible implications for the continued relevancy of correctional programming. A review of the literature identified trends in four key areas which may impact the continued relevancy of correctional programs: (a) education and learning ability; (b) security level designations; (c) sentence lengths; and (d) mental health.

Education and Learning Ability

The majority (78%) of male federal admissions between 1995 and 2004 had not completed high school (Boe, 2005). Of these offenders, 89% were under 25 years of age, of these, 55% had not completed grade 10, and 19% had not completed Grade 8. Similar trends are seen among male Aboriginal offender admissions. Between 1994 and 2000, 87% had no high school diploma, 63.6% had less than Grade 10, and 27% had less than a Grade 8 education (Brown & Motiuk, 2005). These trends have direct implications for correctional programming, in that offenders must have a functional literacy level of Grade 8 as a minimum education requirement for participating in programs (CSC, 2004).

Given the above trends, analyses examined correctional outcomes for two sub-groups of offenders, those having less than a Grade 8 education at federal admission to custody, and those assessed as having learning disabilities. Analyses revealed that low levels of educational attainment and learning disabilities reduce the overall impact of correctional programs on community outcomes, although significant differences remained for certain programs (e.g., living skills programs) and across different community outcomes (see Tables 69 and 70).

Table 63: Community Outcomes for a Sub-sample of Offenders with Less Than Grade 8 Education: Male Offenders
 Offenders With Less Than a Grade 8 Education
   VOP  FVPP  NSAP  SOP(a)  SOP(b)  LSP
 Community Outcomesa
 Any Return  0.85 ns  0.85 ns  0.90 ns  0.89 ns  0.74 ns  0.79***
 Readmission with a new  offence  0.73 ns  0.81 ns  0.76 ns  0.48*  0.60 ns  0.78**
 Readmission with a new  violent offence  0.62 ns  1.37 ns  0.84 ns  0.52 ns  0.57 ns  0.61**
 Readmission with a new sex  offence  ––  ––  ––  0.46 ns  0.44 ns  ––

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01. All figures presented are in relation to a sub-group of offenders, drawn from the comparison group, who were identified as having less than a Grade 8 education. Numbers in italics indicate the data did not adequately fit the model at the 0.1 level of significance.

Table 64: Community Outcomes for a Sub–sample of Offenders Identified as Having Learning Disabilities: Male Offenders
 Offenders Identified as Having Learning Disabilities
   VOP  FVPP  NSAP  SOP(a)  SOP(b)  LSP
 Community Outcomes a
 Any Return  1.03 ns  0.68***  0.86 ns  0.80 ns  0.37***  0.67***
 Readmission with a new  offence  0.73 ns  0.64*  0.83 ns  0.54 ns  0.34*  0.68***
 Readmission with a new  violent offence  0.41*  0.55 ns  0.44**  0.86 ns  0.26**  0.59***
 Readmission with a new sex  offence  ––  ––  ––  2.65 ns  0.25 ns  ––

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01. All figures presented are in relation to a sub-group of offenders, drawn from the comparison group, who were identified as having a learning disability. Numbers in italics indicate the data did not adequately fit the model at the 0.1 level of significance.

Increase in Rates of Maximum Security Classification

For non–Aboriginal male offenders, there is an increasing trend for maximum security designation at admission, from 6% in 1996/97 to 13% in 2005/06, and Aboriginal male offenders are also more likely to be incarcerated in higher security institutions (CSC, Research Branch, 2006; Boe, Nafekh, Vuong, Sinclair, & Cousineau, 2003). In 2004, 27% of offenders were in maximum security settings, representing an increase of 50% since 1997. Although the majority (53%) of women offender admissions initially receive minimum security designations, there has been an increase in maximum security designations since 1996 (from 4% to 11%; CSC, Research Branch, 2006). Increases in higher security designations imply that these offender sub-populations are higher risk, and also confirm the continued relevance of programming offered in higher security institutions. However, because security classification is not used to calculate static risk for Aboriginal or women offenders, it is difficult to determine the direct implications for programming for these sub-populations.

As discussed earlier in this report, there was a large pool of candidates who met the referral criteria for a specific violent offender program who were either not referred to the program, or who were referred and did not enrol. Furthermore, these offenders were more likely to have a maximum designation than their comparison group. Thus, this finding highlights the need for programming offered in higher security institutions.

Shorter Sentences

The proportion of non-Aboriginal male offenders serving less than three years increased by 21% from 1996/97 to 2005/06, with half of new admissions being for sentences of less than three years (CSC, Research Branch, 2006). The proportion of offenders who are in custody serving life/indeterminate sentences also continues to increase, leading to a gradual widening of the gap in sentence length (CSC, Research Branch, 2006). A similar trend is seen among incarcerated Aboriginal male offenders, with 25% serving less than three years (an increase of 9% since 1997), and 21% serving life/indeterminate sentences (a 17% increase from 1997; CSC, Research Branch, 2004). An even greater proportion of women offenders are serving less than three years, with 36% of incarcerated women, and 56% of new admissions, serving shorter sentences (Boe et al., 2003; CSC, Research Branch, 2006). Unlike their male counterparts, the proportion of women serving life/indeterminate sentences decreased from 1997 to 2004, from 21% to 17% (CSC, Research Branch, 2004).

A shorter sentence implies the need for timelier access to programs, and creates challenges in incorporating correctional programs into an offenders correctional plan while keeping in mind his/her earliest possible eligibility dates. Simultaneously, an increase in longer sentences for male offenders could also require more tailored case management approaches and timely incorporation of correctional programming for long-term offenders.

Findings earlier in this report showed that offenders who met the referral criteria for a specific program and were not referred were more likely to be serving shorter sentences (see Findings). As such, this finding highlights the need for an efficient correctional program referral process, as well as correctional programs which can be offered effectively within the time constraints of a short sentence.

Mental Health

The proportion of incarcerated non-Aboriginal male offenders presenting with mental health problems at admission increased from 7% in 1997 to 12% in 2006. In addition, 19% of male offenders in custody are prescribed medication, representing an increase from 9% in 1997 (CSC, Research Branch, 2006). One in ten incarcerated Aboriginal male offenders had a mental health diagnosis in 2004, representing a 67% increase from 1997. Moreover, 17% were prescribed medication an 89% increase from 1997 (CSC, Research Branch, 2004). Relative to their male counterparts, women offenders are more likely to present with a mental health problem. This proportion increased from 13% in 1997 to 21% in 2006 (CSC, Research Branch, 2006).

Given the above trends, analyses examined correctional outcomes for offenders who were diagnosed with a mental disorder at the time of their federal admission. Analyses revealed that, within this group of offenders, program participation remained linked with positive community correctional outcomes for VOPs and living skills programs, while results for FVPPs and SAPs ceased to be significant. Results varied by outcome measure for SOPs and SAPs (see Table 71).

Mental health needs may be potential barriers to correctional intervention, particularly in relation to responsivity, affecting an offenders level of functioning, his or her ability to participate in group settings, and compliance with treatment. Treatment and medication side-effects can further affect an offenders chances of benefiting from a correctional intervention.

Table 65: Community Outcomes for a Sub-Sample of Offenders Diagnosed with Having a Mental Disorder: Male Offenders
 Offenders Identified as Having Mental Disorder
   VOP  FVPP  NSAP  SOP(a)  SOP(b)  LSP
 Community Outcomes a
 Any Return  0.66***  0.81 ns  0.95 ns  0.56*  0.50**  0.77***
 Readmission with a new  offence  0.51***  0.91 ns  0.64*  0.43 ns  0.37 ns  0.82*
 Readmission with a new  violent offence  0.36***  0.86 ns  0.71 ns  1.18 ns  0.15*  0.59**
 Readmission with a new sex  offence  ––  ––  ––  2.58 ns  0.11 ns  ––

Note: a Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01. All figures  presented are in relation to a sub-group of offenders, drawn from the comparison group,  who were diagnosed with having a mental disorder. Numbers in italics indicate the data  did not adequately fit the model at the 0.1 level of significance.

 

FINDING 30: Staff largely perceive existing correctional programs as being responsive to the current offender profile, with the exception of offenders with mental health diagnoses and special needs such as cognitive impairments and low educational attainment. Staff from multi-level security institutions rated the responsiveness of correctional programs higher than staff from other security level institutions.

On average, when asked whether existing correctional programs are responsive to various subpopulations of the current offender profile, one-way within-subjects ANOVA and post-hoc paired t-tests revealed that staff rated their agreement significantly lower for offenders with mental health diagnoses, relative to other offender subpopulations, (F(7, 1015) = 69.01, p < .0001). Agreement as to the ability of programs to respond to offenders with cognitive impairments was rated significantly lower on average when compared to other offender groups, with the exception of those with acute mental health diagnoses. Further, average agreement as to the responsiveness of programs to gang-affiliated offenders was also rated significantly lower on average when compared to other offender groups, with the exception of those with cognitive impairments or acute mental health diagnoses. Agreement as to the ability of correctional programs to respond to offenders with histories of violence was rated significantly higher than that for all other offender groups. Distributions and means for each offender subpopulation are presented in Table 72.

Table 66: Ratings of staff agreement as to whether existing correctional programs are responsive to various offender subpopulations
 Extent of Agreement % (N)
   Strongly Disagree/Disagree  Somewhat Disagree  Somewhat Agree  Agree/Strongly Agree  M  SD
 Subpopulation:
 Acute mental  health  diagnoses  
59.5 (125)
 
15.7 (33)
 
10.5 (22)
 
14.3 (30)
 
2.48
 
1.49
 Gang–affiliated  
41.7 (83)
 
18.6 (37)
 
20.6 (41)
 
19.1 (38)
 
3.06
 
1.46
 History of  violence  
3.5 (8)
 
9.2 (21)
 
21.1 (48)
 
66.2 (151)
 
4.68
 
1.00
 Sentence < 3  yrs  
18.8 (42)
 
22.0 (49)
 
24.7 (55)
 
34.5 (77)
 
3.77
 
1.30
 Young offenders  (< 30 yrs old)  
8.4 (17)
 
11.9 (24)
 
29.7 (60)
 
50.0 (101)
 
4.31
 
1.18
 Dangerous  Offenders  
19.4 (39)
 
21.4 (43)
 
22.9 (46)
 
36.4 (73)
 
3.79
 
1.34
 Serving life  sentences  
14.3 (32)
 
16.5 (37)
 
24.6 (55)
 
44.7 (100)
 
4.08
 
1.25
 Cognitive  impairments  
45.0 (104)
 
19.5 (45)
 
18.2 (42)
 
17.3 (40)
 
2.94
 
1.45

 

Between–groups ANOVAs revealed significant omnibus regional differences in average agreement ratings regarding the ability of correctional programs to respond to gang-affiliated offenders (F(4,193) = 4.09, p < .005), those serving sentences less than three years (F(4,217) = 5.71, p < .0001), and those serving life sentences (F(4,218) = 4.02, p < .005). Post–hoc comparisons further showed that staff from the Ontario Region rated their agreement higher than those from the Quebec Region when asked about gang-affiliated offenders (Mean difference = 1.22, p < .001); staff from the Pacific Region rated their agreement significantly higher than staff from the Prairie Region when asked about offenders serving less than three years (Mean difference = 1.01, p < .005), and those from the Atlantic Region rated their agreement significantly higher than staff from both the Prairie (Mean difference = 1.11, p < .0001) and Quebec Regions (Mean difference = 0.69, p < .05) when asked about the same offender group. Regarding the ability of correctional programs to respond to offenders serving life sentences, staff from the Atlantic Region rated their agreement significantly higher than those from the Pacific (Mean difference = 0.85, p = .005), Prairie (Mean difference = 0.77, p < .05), and Quebec Regions (Mean difference = 0.71, p < .05).

There was a significant main effect of security level when examining differences in average agreement as to the ability of correctional programs to respond to offenders with mental health diagnoses (F(5,200) = 2.44, p < .05), offenders serving sentences less than three years (F(5,213) = 4.12, p = .001), dangerous offenders (F(5,191) = 4.10, p = .001), and offenders serving life sentences (F(5,215) = 3.59, p < .005). Specifically, post-hoc comparisons revealed that staff from multi-level security institutions rated their agreement significantly higher than those from medium security institutions when asked about the ability of correctional programs to respond to offenders with acute mental health diagnoses (Mean difference = 0.98, p < .05). Staff from multi-level institutions also rated their agreement higher than staff from maximum security institutions (Mean difference = 1.12, p < .05), Community Parole Offices (Mean difference = 1.10, p < .05), and regional offices (Mean difference = 1.05, p < .005) when asked about the ability of correctional programs to respond to offenders serving sentences less than three years. Staff respondents from regional offices rated their agreement significantly higher than staff from medium (Mean difference = 1.28, p =.005), maximum (Mean difference = 1.50, p < .05), and multi–level security institutions (Mean difference = 1.40, p < .01) when asked about the ability of correctional programs to respond to Dangerous Offenders. Lastly, staff from minimum security institutions rated their agreement significantly higher than those from medium security institutions (Mean difference = 0.79, p < .05) and regional offices (Mean difference = 1.10, p < .05) when asked about the ability of correctional programs to respond to offenders serving life sentences.

 

FINDING 31: Almost all staff felt that more could be incorporated into correctional programming that would address the needs of offenders with mental health needs. The need for new programs for those with mental health needs and one-on-one intervention opportunities were commonly mentioned by staff.

Staff members were asked whether there was anything specific to offenders with mental health concerns that should be incorporated into correctional programs that would increase their likelihood of successful reintegration into the community. Of the 203 respondents who answered definitively (yes or no), 183 (90.1%) indicated that there were areas where specific changes could be made for offenders with mental health needs, and 186 (91.6%) went on to make recommendations for change. The most common recommendations included the need for new programs for those with mental health needs (38.2%; N = 71), the need for more one-on-one interventions and psychological services (16.1%; N = 30), and the need for more staff training and skill development related to mental health needs (12.4%; N = 23). Individual recommendations endorsed by more than 5% of respondents are listed in Table 73, with a complete list of recommendations listed by staff members provided in Appendix I, Table I5.

Table 67: Summary of Recommendations for Improving Correctional Program Outcomes for Offenders with Mental Health Needs
 Recommendations:  Staff Respondents N (%)
 Create new programs for those with mental health needs  71 (38.2%)
 More 1–on–1 interventions/psychological services  30 (16.1%)
 More training/skills for staff  23 (12.4%)
 Adjust currents programs to accommodate needs  22 (11.8%)
 More resources (funding, psychologists)  22 (11.8%)
 More identifying assessments  16 (8.6%)
 More community support/follow–up  16 (8.6%)
 More team–oriented approach  13 (7.0%)
 More communication regarding offender mental health  11 (5.9%)
 More flexible program delivery schedule  10 (5.4%)

Note: The recommendations listed make up 83.9% of those mentioned by staff members.

FINDING 32: Almost all staff felt that more could be incorporated into correctional programming that would address the needs of offenders with special needs such as cognitive impairments and learning disabilities. The development of new programs and the need to adapt current programming to meet these special needs were stressed by staff.

Staff members were asked whether there was anything specific to offenders with special needs (i.e., cognitive impairments, learning disabilities, etc) that should be incorporated into correctional programs that would increase their likelihood of successful reintegration into the community. Of the 208 respondents who answered definitively (yes or no), 183 (88.0%) indicated that there were areas where specific changes could be made for offenders with special needs, and 194 (93.3%) went on to make recommendations for change. The most common recommendations included the need for new programs specific to the special needs exhibited by these offenders (29.4%; N = 57), the need to adapt current programming to meet special needs (25.3%; N = 49), and the need for more staff training related to special needs (20.6%; N = 40). Individual recommendations endorsed by more than 5% of respondents are listed in Table 74, with a complete list of recommendations listed by staff members provided in Appendix I, Table I6.

Table 68: Summary of Recommendations for Improving Correctional Program Outcomes for Offenders with Special Needs
 Recommendations:  Staff Respondents N (%)
 Implement new programs specific to group/need  57 (29.4%)
 Adapt current programs to meet needs  49 (25.3%)
 More staff training  40 (20.6%)
 More 1–on–1 programming  21 (10.8%)
 More tutors/teacher assistants/specialists  21 (10.8%)
 More flexible program schedules  20 (10.3%)
 More identifying/assessment tools  19 (9.8%)
 More resources  19 (9.8%)

Note: The recommendations listed make up 82.3% of those mentioned by staff members.

 

FINDING 33: The majority of staff members indicated that the implementation of correctional programs might be impacted both positively and negatively, should legislation eliminating statutory release come into effect. Potential impacts include both increasing and decreasing offenders motivation to participate in programs, increased waitlists for programs and a greater demand for programs in institutions.

The majority of staff (93.5%, N = 216) indicated that the potential elimination of statutory release might impact correctional programs. Specifically, respondents suggested that offenders will be more motivated to participate in programs (48.8%, N = 104), that offenders will be less motivated to participate in programs (26.3%, N = 56), that offenders will have a longer amount of time in the institution in which to participate in programs (24.9%, N = 53), and there will be a greater demand / need for correctional programs in institutions (17.4%, N = 37). A complete list of potential implications of eliminated statutory release is presented in Appendix I, Table I7.

RECOMMENDATION 4:

CSC should develop a strategy that addresses the programming needs of the increased proportion of offenders with education and learning deficits, mental disorders, shorter sentences, and higher security level ratings.

3.7 Objective 7: Unintended Findings: Unintended findings are areas wherein the initiative created or encountered any positive or negative effects.

 

FINDING 34: The majority of staff indicated that more could be incorporated into correctional programming that would address the needs of Aboriginal offenders. The need for a greater number and a greater variety of Aboriginal-specific programming was stressed, as well as the need to incorporate Aboriginal content into existing correctional programs.

Staff members were asked whether there was anything specific to Aboriginal offenders that should be incorporated into correctional programs that would increase their likelihood of successful reintegration into the community. Of the 199 respondents who answered definitively (yes or no), 155 (77.9%) indicated that there were areas where Aboriginal-specific changes that could be made, and 160 (80.4%) went on to make recommendations for change. The most common recommendations included the need for more (and a greater range of) Aboriginal specific programming (37.5%; N = 60), the need for more Aboriginal content in existing correctional programming (16.9%; N = 27), and increasing the amount of resources that go toward Aboriginal programming, specifically in hiring more Elders, Aboriginal Liaison Officers (ALO), and therapists (14.4%; N = 23). Individual recommendations endorsed by more than 5% of respondents are listed in Table 75, with a complete list of recommendations listed by staff members provided in Appendix I, Table I8.

Table 69: Summary of Recommendations for Improving Correctional Program Outcomes for Aboriginal Offenders
 Recommendations:  Staff Respondents N (%)
 More/greater range of Aboriginal specific programming  60 (37.5%)
 More Aboriginal content in existing programming  27 (16.9%)
 More resources (Elders/ALOs/therapists)  23 (14.4%)
 More staff training on Aboriginal culture  19 (11.9%)
 More Aboriginal staff (CPOs/POs)  18 (11.3%)
 More contact with facilitators/Elders  15 (9.4%)
 More community liaising/connections  15 (9.4%)
 Greater use of healing plans/Elder assessments  13 (8.1%)
 More structure/accreditation in Aboriginal programs  8 (5.0%)
 More Elder training on correctional approach/reports  8 (5.0%)

Note: The recommendations listed make up 78.6% of those mentioned by staff members.

 

FINDING 35: The majority of staff indicated that more could be incorporated into correctional programming that would address the needs of women offenders. Specifically, the need for a greater number and a greater variety of women-specific programming was expressed by almost half of staff respondents.

Staff members were asked whether there was anything specific to women offenders that should be incorporated into correctional programs that would increase their likelihood of successful reintegration into the community. Of the 154 respondents who answered definitively (yes or no), 56 (63.6%) indicated that there were areas where women-specific changes could be made, and 60 (40.0%) went on to make recommendations for change. The most common recommendations included the need for more (and a greater range of) women-specific programming (46.7%; N = 28), the need for more contact with the community and ETAs (10.0%; N = 6), the need for more community programming (10.0%; N = 6), the need for more consideration of motherhood in program content (10.0%; N = 6). Individual recommendations endorsed by more than 5% of respondents are listed in Table 76, with a complete list of recommendations listed by staff members provided in Appendix I, Table I9.

Table 70: Summary of Recommendations for Improving Correctional Program Outcomes for Women Offenders
 Recommendations:  Staff Respondents N (%)
 More/greater variety of women specific programming  28 (46.7%)
 More community contact/ETAs  6 (10.0%)
 More community programs  6 (10.0%)
 More consideration of motherhood in programs  6 (10.0%)
 Treat them the same as male offenders  4 (6.7%)
 Improve women’s risk assessments  4 (6.7%)
 More staff/resources  3 (5.0%)
 More vocational programs/training  3 (5.0%)

Note: The recommendations listed make up 76.9% of those mentioned by staff members.

 

4.0 APPENDICES

Appendix A: Detailed Program Descriptions


Pre-Program Interventions

The Motivation-based Intervention Strategy (MBIS) is a pre-program intervention implemented in segregation units88 of maximum security institutions (except one) and at the Special Handling Unit. The intervention strategy, for offenders exhibiting disruptive behaviours, addresses offenders motivation and prepares them for transfer to the general population. The strategy is comprised of direct intervention and skills workshops (on topics such as self-monitoring, anger control, high-risk thinking, communication, slips and craving management, problem-solving) designed to assist inmates to increase their motivation to change and to engage in their correctional plans. Although it is not a replacement for correctional programs, MBIS uses the same research-based techniques and principles that make CSCs correctional programs effective (Blanchette & Moser, 2006).

 

Family Violence Prevention Programs

High Intensity Family Violence Prevention Program

This program is designed for federally-sentenced male offenders who have demonstrated a pattern of violence against intimate female partners, have committed at least two confirmed acts of violence against one or more intimate partners, and have been assessed as high risk on the SARA, a 20-item clinical checklist of risk factors for spousal assault. The program consists of approximately seventy-five 2.5-hour group sessions and is only offered in institutions. Each individual also has 8-10 individual counselling sessions.

Moderate Intensity Family Violence Prevention Program

To be enrolled in this program, a male offender must have committed at least one confirmed act of family violence against an intimate partner and must present moderate risk based on their SARA results. The program consists of approximately twenty-six 2.5-hour group sessions delivered over a period of 5-13 weeks. Each participant also has 3 individual counselling sessions. The program is offered in institutions and in the community.

Aboriginal High Intensity Family Violence Prevention Program

This program is for Aboriginal male offenders who are rated as high risk on the SARA and who have a documented history of two or more incidents of violence against a female partner. The objectives of the program are to: eliminate violence and abuse toward female partners and in the family context; develop insight, relationship skills, thinking skills, and emotion management skills to promote healthy relationships; accept responsibility; increase awareness of unhealthy (abusive) attitudes and behaviour towards women partners and children; and reduce the intergenerational transmission of violence. This program consists of 72 sessions, including 2.5-hour group sessions and full day ceremonies. Ten sessions are conducted individually with the program facilitator, and individual sessions with the Elder are offered.

 

Substance Abuse Programs

National Substance Abuse Program High Intensity

This program is for offenders who demonstrate a significant or serious need based on the results of their computerised assessment and those offenders whose crimes are related to substance abuse. The NSAP-High is offered in institutions and consists of 89 2-hour sessions.

National Substance Abuse Program Moderate Intensity

This program is for offenders who demonstrate a moderate need based on the results of their computerised assessment and whose crimes are related to substance abuse. This program is offered in institutions. It consists of twenty-five 2-hour group sessions and one individual session.

National Substance Abuse Program Low Intensity

This program is for offenders who demonstrate a low need based on their computerised assessment and whose crimes are related to substance abuse. The NSAP-Low consists of 10 sessions approximately 2.5 hours in length, followed by 12 weekly maintenance sessions.

Aboriginal Offender Substance Abuse Program

This program is for Aboriginal male offenders who demonstrate a need for a high intensity program, as indicated by scores on the Alcohol Dependence Scale (ADS) and/or the Drug Abuse Screening Test (DAST) and/or Problems Related to Drinking Scale (PRD). The program objectives include the following: develop awareness of linkages between substance abuse and criminality; develop opportunities to engage and motivate participants in positive change process; develop a skill base to promote an abstinence lifestyle in the four domains of the Circle (physical, mental, emotional and spiritual); and introduce and incorporate spiritual and cultural activity as a means of healing and living a healthy lifestyle. The program consists of 33 sessions.

Women Offender Substance Abuse Program

This program is designed for women offenders who demonstrate a moderate to high risk in the substance abuse domain. It is offered in institutions, and consists of 100 sessions, each lasting 2.5 hours. The program offers intervention and support from admission through to warrant expiry by recognizing that needs are dynamic and that intervention is most effective when it reflects context as well as content.

 

Violent Offender Programs

Violence Prevention Program

Because the targets of change are usually complex and multiple, the VPP is an intensive intervention that integrates a variety of rehabilitative approaches. VPP is for offenders who are considered at high risk to commit violent crimes and for those offenders who have already committed at least two violent offences and/or murder. The program is offered in institutions only and consists of 94 two-hour group sessions and four individual counselling sessions.

In Search of Your Warrior Program

This program is for Aboriginal male offenders who have a history of violent offending and who are considered a high risk to re-offend violently. Participants must be actively following native spirituality. The program requires a high degree of commitment from the participant and includes a number of traditional ceremonies. The program is four months in length and is offered in institutions.

Spirit of A Warrior Program

This program is for Aboriginal women offenders who have a moderate to high need for treatment of violence. The program consists of an in-depth intervention that is intended to reduce the risk to re-offend with violence, reduce risk to relapse, improve family relations, improve the ability to communicate with others, improve coping skills, and adapt Aboriginal culture and spirituality into all aspects of behaviour and everyday life. It is expected that with a more informed base of traditions, Aboriginal women will be better able to manage their lives. The program is offered in institutions and consists of 92 group sessions, each lasting 1 to 2 hours.

 

Sex Offender Programs

National Moderate Intensity Sex Offender Program

Participants must demonstrate a moderate risk of committing another similar offence, as indicated by scores on the SSOA and the Static-99, and have moderate or high criminogenic need, as indicated on the Stable-2000. The program is available only in institutions and is between 160 and 280 hours in length and has two individual counselling sessions.

National Low Intensity Sex Offender Program

Participants must present a low risk for committing another similar offence, and have low or moderate criminogenic needs. The program is available in institutions and is between 24 and 60 hours in length and has two individual counselling sessions.

Sex Offender Treatment for Women

This program is for women offenders who have been convicted of a sex offence or a sexually motivated crime, or those who have admitted to committing a sex offence for which they received no convictions. The program focuses on five key areas: self management; deviant sexual arousal; cognitive distortions; intimacy, relationships and social functioning; and empathy and victim awareness. The primary goals of treatment for women who sexually offend are to learn to identify the factors that influenced their offences, and to learn how to deal more effectively with them in order to reduce the risk of re-offending. Women participate in sessions once or twice a week. This program is available in institutions and the community and lasts 3 to 4 months, or up to a maximum of 70 sessions.

 

Life Skills Programs

Reasoning and Rehabilitation Program

R & R is a moderate intensity program that teaches male and female offenders skills that address the socio-cognitive deficits linked to criminal behaviour. Specific target areas include self-regulation and self-management, egocentrism/lack of social perspective, problems related to assertiveness and problematic social interactions, poor critical reasoning, and rigid thinking. The program consists of 37 group sessions each lasting 2 to 3 hours and four individual sessions. It is offered in institutions and in the community. Following the program, male and female offenders who still have significant socio-cognitive deficits are directed to the Cognitive Skills Maintenance Program.

The Anger and Emotions Management Program

The A & EM program is a moderate intensity program for offenders who are assessed at moderate or high risk to re-offend, or who have low to moderate reintegration potential. Offenders must also have criminal behaviour associated with inappropriate management of their emotions, some or considerable need in the personal/emotional domain, and score positively on at least one of the following indicators:

  • Aggression
  • Assertion problem
  • Poor conflict resolution
  • Hostility problem
  • Poor stress management
  • Low frustration tolerance
  • Worries unreasonably
  • Problematic risk taking
  • Thrill seeking

A & EM is offered to male and, by using an adapted version, women offenders. The program teaches offenders to change their thinking patterns that trigger emotions associated with their criminal behaviour. The program consists of 27 group sessions, each lasting 2 to 3 hours, and also includes two individual sessions. It is available in institutions and in the community. Following the program, offenders who still have problems are directed to the Anger and Emotions Management Maintenance Program.

Counter-Point Program

Counter-Point is a moderate intensity community-based program offered through larger parole offices in all regions. The principal goal of the program is to help participants change their pro-criminal values and attitudes and take more responsibility for their criminal actions. The program focuses on the following themes: (a) setting goals to lead a crime-free life; (b) identifying tools that can help bring about the desired change; (c) identifying the thinking patterns conducive to criminal behaviour; (d) identifying a pro-social network; (e) solving interpersonal problems; and (f) developing a relapse prevention plan. This program is designed for male offenders on conditional release (including temporary absences) who have a moderate to high risk of re-offending or a low to moderate reintegration potential. The program consists of 20 group sessions 2 to 3 hours in length and 5 individual sessions.

Circles of Change Program

This program is for Aboriginal women offenders who demonstrate a moderate to high need in the associates/social interaction domain and/or the general attitude domain, as well as those who have a demonstrated deficit in problem-solving. This cognitive-behavioural and solution-based program teaches participants about Canadian Aboriginal culture, communication styles, self-esteem and self-care issues, problem-solving skills, womens roles in their family of origin, healthy and unhealthy relationships, and social injustice. The program is offered in institutions and consists of 19 modules delivered in 25 to 30 sessions, each lasting 2.5 hours.

National Aboriginal Basic Healing Program

This is a moderate intensity program for male Aboriginal offenders who are actively participating in Aboriginal spiritual practices. The program develops a spiritual foundation. Other program objectives are: (a) acquiring historical knowledge about Aboriginal people, the land, and colonization; (b) healing from personal childhood traumas; (c) developing effective communication and problem solving skills; and (d) learning about the cultural and spiritual traditions, thereby changing attitudes and behaviours. If the offender is referred to this program and another national Aboriginal program, he is expected to take this one first. The program contains 60 sessions.

 

Maintenance Programs

Community Maintenance Program (Generic Model)

This maintenance program is designed to follow most national programs and is delivered to offenders who present a moderate to high risk to re-offend. The program reviews a selected sub-set of core skills learned in the national programs such as (a) working towards goals, (b) dealing with emotions, (c) countering high risk thinking, (d) interpersonal communication, and (e) self-management. The program consists of 12 sessions, each lasting 2 hours.

Family Violence Maintenance Program

The program is intended for offenders who have completed the high or moderate intensity family violence programs. Participants in institutions are required to attend maintenance for at least six months prior to release to the community. For community maintenance, offenders are required to attend at least six sessions, at which point their risk will be re-evaluated. Maintenance focuses on their current situation and a review of their relapse prevention plan.

Sex Offender Maintenance Program

Sex offenders can participate in the Sex Offender Maintenance Program for as long as required by their level or risk or need. The length of maintenance is based on the needs of each offender. Maintenance is available in both the institution and the community.

National Substance Abuse Maintenance Programs

All offenders who have completed a SAP must participate in maintenance. The frequency of participation is based on risk and need. Offenders are discharged from maintenance once they have demonstrated a period of stability. They can be re-referred to maintenance at any time should they lapse.

Cognitive Skills Maintenance Program

This program is for offenders who completed the R & R program or the Cognitive Skills Program. Participants must demonstrate the need to continue practicing the skills learned in the Cognitive Skills Maintenance Program. This program can consist of one intervention of at least 10 sessions, depending on the offenders needs. Each session is two to three hours in length. It is offered in institutions and in the community, and can be delivered to open groups (i.e., continuous in and out) or closed groups (i.e., to the same group of offenders from the beginning to the end of the program).

Anger and Emotions Management Maintenance Program

This program is for male and female offenders who have finished the A & EM and who still demonstrate the need to practice skills learned in this program in order to obtain an impact on the associated risk factors. The program consists of eight sessions lasting 2 to 3 hours each. It is offered in institutions and in the community.

Relapse Prevention / Maintenance Program for Women Offenders

This program is for women offenders who demonstrate a moderate to high need of intervention. It consists of 20 group sessions, each lasting 2.5 hours. The program is intended to sustain and enhance program gains as well as function as a community program for women, based on the risk they represent and is dynamic enough to meet the needs of all women. The program targets lifestyle factors that are contributing factors in increasing the risk that women face while under community supervision. The program is offered in the institution and in the community. Women offenders may retake the program as needed, and there is no pre-requisite for participation.

Appendix B: Evaluation Matrix

 Key Results  Questions  Analyses*  Information Sources
 Objective #1: Continued Relevancy Do Correctional Programs remain and continue to  remain consistent with departmental and government-wide priorities?
 1. The ability of  programs to  address the risks  and needs of  offenders has  changed in  accordance with  the changing  offender profile
  • How responsive have/will programs be to changes in the offender profile?
  • Review of past, current, and future offender profile trends that impact risk, need and responsivity.
  • Relationship between changing demographic and program assignments overtime
  • Relationship between changing demographic and program development -ability to change program structure and content to meet new demographic needs
  • Review of human resource staffing/ training plan or succession planning
  • CSC’s automated OMS.
  • Key informant interviews
  • Research reports
 2. Correctional  programs will be  responsive to  legislation that  impacts the  demographic  profile of the  federal offender  population
  • How responsive can correctional programs be to changes resulting from anticipated legislation?
  • Review of upcoming legislation impacting the profile of the federal offender population.
  • Identification of additional programming needs, and capacity to meet those needs, for new demographic (if any).
  • Key informant input Department of Justice (DOJ)
  • OMS data number of offenders who have been convicted of crimes that would be affected by upcoming legislation.
  • Key informant interviews
 Key Results  Questions  Analyses  Information Sources
 Objective #2: Implementation: Have correctional programs been managed in such a  way that goals and objectives can be realistically achieved, and have management  implementation issues been adequately considered?
 1. Offenders are  receiving  appropriate  programs
  • Are offenders with identified needs being referred to the recommended programs?
  • Are offenders with identified needs assigned the recommended program?
  • Compare offenders overall dynamic need levels with the program interventions referred to, to see if there is congruency in intensity level.
  • Investigate how many offenders are being assigned to the recommended program intervention.
  • OMS
 2. The current  correctional  program  accreditation  process is  relevant
  • Is the accreditation process feasible?
  • Is the accreditation process relevant for Women and Aboriginal offender programs?
  • Review of accreditation process
  • Outcomes analysis (institutional and community)
  • Key informant input
  • OMS
  • CPIC
  • Accreditation documentation

___________________________________

* All analyses were disaggregated by: (a) male non-Aboriginal offender programs;
(b) Aboriginal offender programs; and (c) women offender programs, unless otherwise stated.

 Key Results  Questions  Analyses  Information Sources
 Objective #3: Success (Effectiveness) Are Correctional Programs89 meeting the  intended results?
 1. Belief and  attitude  changes are  achieved  through  program  participation
  • Do correctional programs improve beliefs and attitudes toward treatment targets?
  • Are belief and attitude changes associated with long-term outcomes?
  • Pre-/Post-examination of offenders attitudes for the programs of interest.
  • Model the relationship between belief/attitude change with institutional and community outcomes (see Key Results 2 and 3).
  • OMS
  • Pre–/Post-measures
 2. Positive  institutional  outcomes are  achieved
  • Does program participation lead to a reduction in the involvement of major and minor institutional incidents90 (perpetrator, associate)?
  • Are those who participate in programs deemed to be manageable in the community, and thus more likely to be granted a discretionary release by the NPB?
  • Compare institutional incidents of program participants with a matched cohort91 of offenders assigned and not yet received the program. A multiple regression model will model the relationship between program interventions (independent variable) with institutional incidents. Institutional incidents will be grouped into major, minor, and substance-abuse related (including urinalysis test results) groupings.
  • Comparison of release dates and release types of program participators with offenders assigned to programs but who have not participated. Response variables for modelling will include proportion of sentence served until release and type of release. Independent variable will be program participation (yes / no).
  • OMS
 3. Positive  community  outcomes are  achieved
  • Are offenders who have taken programs that address their needs: (a) more likely to stay in the community longer; (b) less likely to return with a technical revocation; (c) less likely to return with a new offence; (d) more likely to reach the end of their sentence; or (e) more likely to remain in the community past the end of their sentence?
  • Compare offenders who participated in a particular program with a matched cohort of offenders who have been assigned to that particular program but have not yet received the program.
  • Survival Analysis of the released cohort.
  • OMS
  • CPIC

 

 Key Results  Questions  Analyses  Information Sources
 Objective #3: Success (Efficiency) Is the strategy producing its planned outputs in  relation to expended resources?
 1. Streamlined  pre-post- program  assessment  batteries have  the ability to  predict future  success.
  • Is there evidence to suggest that the use of every assessment battery test is linked to success or would a shorter version/ a single program-specific version or a single generic version (GPPM) be more efficient?
  • Improvements made on: (a) pre-/post-program assessment batteries (multiple and/or single program-specific batteries); and/or (b) a single generic pre-/post-assessment (GPPM) are correlated with institutional and community outcomes.
  • All program pre– and post–program assessment batteries.
  • GPPM data.
  • CPIC and OMS follow-up data on program participants
 2. There are  sufficient  numbers of  participants in  programs in  order for them  to be delivered
  • Are the number of offenders participating in each cycle of the correctional program within the stated limits (i.e., not too few or to many offenders participating in each delivery of the program)?
  • Review the number of enrolments and program cancellations in each cycle of the program to determine if they are within the limits prescribed.
  • Review of program guidelines
  • Programs database number of enrolled participants and program cancellations in each cycle of the program
 3. There are  sufficient  numbers of  trained CPOs  given the  required  number as per  group setting.
  • Are there enough CPOs and Aboriginal CPOs to deliver the required menu of correctional programs at each site?
  • Review of Human Resource Management Sector (HRMS) and OMS data to identify any gaps where there are limited CPOs to deliver programs.
  • HRMS data number of CPOs trained, number certified, time between training and certification
  • HRMS data staff retention
  • Key informant input CPOs, Wardens, Regional Managers
 4. There is  sufficient,  efficient, and  appropriate  training for  CPOs.
  • Are training resources appropriately directed toward staff that will, once they have received the training, deliver the program to offenders?
  • With CPOs cross-trained in multiple programs, is the CPO program specific training repetitive?
  • Review of HRMS and OMS to determine how many trained CPOs continue to deliver programs.
  • Review of all national program training schedules to identify overlap in content between program training.
  • Assessment of the associated cost savings of isolating integrated content from correctional programs, and impact on facilitator and/or program. ( i.e. cost analysis of introducing a generic training for integrated content)
  • HRMS data number of CPOs trained to deliver programs, and mean length of time spent delivering program (or mean number of sessions delivered)
  • Key informant input CPOs, trainers
  • Review of CPO training agendas for each correctional program
  • IFMMS data
 Key Results  Questions  Analyses  Information Sources
 Objective #3 (continued) : Success (Efficiency) Are correctional programs producing  planned outputs in relation to expended resources?
 5. Correctional  programs are  delivered to  inmates in the  most efficient  and effective  manner
  • Would there be efficiency gains if integrated program content were removed from individual programs and offered in a single module?
  • Assessment of whether there are time savings by comparing: (a) actual duration of the suite of programs to which offenders are referred; with (b) duration of programs under an alternative scenario, where the time associated to integrated content is removed from the suite of program referrals and offered in a single module.
  • Document review
  • Key informant interviews
  • OMS data
 6. The average  amount of time  spent in a  program is  commensurate  with the  intensity level of  the program
  • Are the number of hours spent in the program commensurate with the number of hours required and stated within Commissioners Directive 726
  • Comparison of hours required for each intensity level and number of hours spent in each program.
  • Program manuals indicating time spent in program
  • Commissioners Directive 726 detailing number of program hours required for each intensity level of program
 Key Results  Questions  Analyses  Information Sources
 Objective #3: Cost-Effectiveness: What is the relationship between the amount  spent and the results achieved relative to alternative design and delivery  approaches?
 1. Correctional  outcomes are  cost-effective
  • What is the effectiveness of correctional institutional and community outcome measures relative to the associated correctional program costs?
  • Associate costs to effectiveness outcomes (i.e. $ per day extended in the community versus costs of incarceration).
  • OMS
  • IFMMS
  • CPIC
 2. Delivery  designs and  approaches are  cost-effective?
  • Are there alternative program designs and delivery approaches that may yield the same or better outcomes at less cost?
  • Associate costs to efficiency outcomes (see Efficiency items #4 and #5).
  • IFMMS

 

Appendix C: Referral Criteria for Correctional Programs and Corresponding Variables in Used in the Approximated Measure


 Program  Referral Criteria  Approximated Referral Measure  Data Source
 Violence Prevention Program  Statistical Information on Recidivism Scale (SIR) rating of poor or very poor; and

 Two or more separate convictions for Schedule I (CCRA) offences and/or homicide

 Statistical Information on Recidivism Scale (SIR) rating of poor or very poor; and

 Two or more separate convictions for Schedule I (CCRA) offences and/or a conviction for homicide

 Offender Management System (OMS)

Canadian Police Information Centre (CPIC) data

 In Search of Your Warrior  Aboriginal male offenders with a history of violent offending; and

 Assessed as high risk to re–offend violently; and

 Actively following native spirituality

 Two or more separate convictions for Schedule I (CCRA) offences and/or homicide conviction; and

 High Overall Static Risk Level Rating

 Not Assessed

 CPIC

 

 OMS

 

 Spirit of A Warrior  Aboriginal women for whom violence has been identified as a moderate to high need  Two separate incidents of violence (two Schedule I offences) resulting in moderate to high degree of harm, and moderate to high risk to re-offend. This excludes offences against children and sex offences. Provincial data were also considered for previous offence history; OMS Offender Intake Assessment (OIA) flags were used to assess harm to victim and Overall Static Risk Level Rating.  CPIC

 OMS

 National Substance Abuse Program – High Intensity  Assessed as having substantial/ severe need on CLAI/CASA; and

 Has an established link between crime and substance abuse

 CLAI/CASA outcome ratings

 

 Not assessed

 OMS
 National Substance Abuse Program – Moderate Intensity  Assessed as having moderate/ intermediate need on CLAI/CASA; and

 Has an established link between crime and substance abuse

 CLAI/CASA outcome ratings

 Not assessed

 OMS
 National Substance Abuse Program – Low Intensity  Assessed as having low need on CLAI/CASA; and

 Has an established link between crime and substance abuse

 CLAI/CASA outcome ratings

 Not assessed

 OMS
 Aboriginal Offender Substance Abuse Program  Aboriginal by birth or a demonstrated commitment to an Aboriginal way of life; and

 High need as demonstrated on Alcohol Dependence Scale (ADS) and/or the Drug Abuse Screening Test (DAST) and/or Problems Related to Drinking Scale (PRD); moderate need offenders also referred as an alternative to NSAP-Moderate

 Self–identifiers of Aboriginal descent

 CLAI/CASA outcome ratings

 

 OMS

 

 OMS

 Women Offender Substance Abuse Program  Need for substance abuse programming  CLAI/CASA outcome ratings  OMS
 High Intensity Family Violence  High Risk Rating on the Spousal Assault Risk Assessment (SARA); and

 A documented history of two or more incidents of violence against a female partner

 High Risk Rating on the Spousal Assault Risk Assessment (SARA); and

 OIA indicator (spousal abuse perpetrator) or a documented history of two or more incidents of violence against a female partner, or item #11 of the SARA (physically assaulted past or current partners)

 OMS

 


 OMS

 CPIC

 Moderate Intensity Family Violence  Moderate risk rating on the SARA; and

 A documented history of one incident of violence against a female partner

 Moderate risk rating on the SARA; and

 OIA indicator (spousal abuse perpetrator) or a documented history of one incident of violence against a female partner, or item #11 of the SARA (physically assaulted past or current partners)

 OMS

 


 OMS

 CPIC

 Aboriginal High Intensity Family Violence Prevention Program  Moderate risk rating on the SARA; and

 A documented history of two or more incidents of violence against a female partner

 High risk rating on the SARA; and

 OIA indicator (spousal abuse perpetrator) or a documented history of two or more incidents of violence against a female partner, or item #11 of the SARA (physically assaulted past or current partners)

 OMS

 


 OMS

 CPIC

 High Intensity Sex Offender Programs  Convicted of a sex offence; or convicted of a sexually motivated crime; or admitted to a sex offence for which they have not been convicted; and

 High risk to re-offend relative to other sex offenders, with moderate to high levels of criminogenic needs

 OIA indicator (current or past sex offending), provincial sentence for a sex offence; and

 

 

 

 “High” priority–rating on an approximated version of the Static 99 and Stable 2007. The Priority Rating is comprised of the combined overall outcomes of these scales (Hanson, Harris, Scott & Helmus, 2007). Approximated items for both the Static 99 and Stable 2007 were taken from a combination of OIA flags and provincial data.

 CPIC

 

 

 

 

 

 OMS

 CPIC

 Moderate Intensity Sex Offender Programs  Convicted of a sex offence; or convicted of a sexually motivated crime; or admitted to a sex offence for which they have not been convicted; and

 Moderate risk to re–offend relative to other sex offenders, with moderate to high levels of criminogenic needs; or high risk to re-offend relative to other sex offenders, with moderate levels of criminogenic needs

 OIA indicator (current or past sex offending), provincial sentence for a sex offence; and

 

 

 

 “Moderate” priority-rating on an approximated version of the Static 99 and Stable 2007. Approximated items for both the Static 99 and Stable 2007 were taken from a combination of OIA flags and provincial data.

 CPIC

 

 

 


OMS
 Low Intensity Sex Offender Programs  Convicted of a sex offence; or convicted of a sexually motivated crime; or admitted to a sex offence for which they have not been convicted; and

 Low risk to re–offend relative to other sex offenders, with low to moderate levels of criminogenic needs

 OIA indicator (current or past sex offending), provincial sentence for a sex offence; and

 

 

 

 “Low” priority–rating on an approximated version of the Static 99 and Stable 2007. Approximated items for both the Static 99 and Stable 2007 were taken from a combination of OIA flags and provincial data

 CPIC

 

 

 

 OMS

 Clearwater Sex Offender Program  Score moderate–high or high on the STATIC–99; and

 Has special needs that preclude them from attending lower intensity institutional programs; and

 Demonstrates persistent deviant sexual preference or other paraphilias, even if these have not led to conviction; and

 Is not suitable from less intensive treatment as a result of extreme minimisation or denial

 OIA indicator (current or past sex offending), provincial sentence for a sex offence; and

 “High” priority–rating on an approximated version of the Static 99 and Stable 2007. Approximated items for both the Static 99 and Stable 2007 were taken from a combination of OIA flags and provincial data.

 OMS

 

Appendix D: Intensity, Duration and Setting of Correctional Programs92

Target Population  

Overall Risk

Needs

Intensity

Average Duration

Group Setting

High
High or Moderate to High
;High
Minimum 15 to 36 weeks Minimum 10 to 15 hours per week Closed or open groups 2 correctional program officers
Moderate – High Moderate to Low

 

Moderate


Between 5 to 25 weeks

5 to 15 hours per week


Closed or open groups

1 or 2 correctional program officers according to the program manual

Moderate Moderate
Moderate – Low Moderate to High
Moderate – Low

Low Or Moderate to Low


Low

Between 1 to 16 months
2 to 6 hours per week

Closed or open groups
1 correctional program officer
Low

Offenders who complete program
Given to moderate and high risk offenders

 

Maintenance
Duration and frequency
determined by the correctional
program officer after discussion
with the parole officer
Open groups

1 correctional program officer

 

Appendix E: Generic Program Performance Measure (GPPM)

The GPPM is a generic pre–/post-program assessment instrument completed by program facilitators to gauge offenders progress toward generic treatment targets. All correctional programs have some common goals such as achieving skills and knowledge gain, changing attitudes (performance), and the extent to which the offender puts forth effort is also important (effort). Unlike other pre–/post-measures, the GPPM is used for multiple programs and is linked through OMS.

The GPPM is composed of three scales that measure: (a) performance (skill, attitude, and knowledge gain); (b) responsivity (factors that could affect the participants progress in treatment); and (c) effort (the extent to which the participant tries to address the objectives of the program through completion of work, class participation, etc.). Items are scored on a 5–point scale (–2, –1, 0, +1, +2), with lower scores indicating an absence (–2) or inconsistent (–1) evidence of the elements of the program objective, and higher scores indicating a good (+1) or excellent (+2) level of performance on the elements of the objective. To calculate a participants score, item scores are summed and averaged.

The pre–treatment or initial GPPM rating should be completed based on the participants early performance in group sessions (no later than the second session) and on information provided through pre–program interviews and file information. The final rating is completed at the end of the program (after the participants last session).

The GPPM has been used to evaluated program participation and performance of offenders with many different correctional programs including violence prevention, family violence, sex offender, substance abuse, and life skills programs. Results will be provided for the intensity level of the program and will be disaggregated for non-Aboriginal male offenders, Aboriginal male offenders and women offenders.

Performance and responsivity scales are recorded pre- and post–program, and a test of differences will be reported for these scales. For the performance scale there is one additional question in post–program. Therefore, it is recommended to use a mean score where the questions are added together and divided by the number of questions. This approach is taken for all three scales of the GPPM.

The effort scale is only recorded at post as it is difficult to gauge effort prior to program participation. Therefore, a mean will be reported for effort as a difference score cannot be calculated. The possible range of these mean scores is –2 to +2. Negative t-test scores (t values) indicate that offenders are having ratings increase at the end of the program as compared to at the beginning of the program. This is what is expected to occur.

 

Results for Non-Aboriginal Male Offenders

Results from all intensity levels of institutional programs and community programs reveal significant changes on both the performance and responsivity scales (p < .05 or greater). These results suggest that facilitators are rating offenders more positively on areas related to program targets after completion of the program.

 

Performance and Responsivity Results for Non–Aboriginal Male Offenders
 Performance and Responsivity Subscales
 Intensity Level  Scales  n  t  p
 Low Intensity Programs  Performance  149  -15.15  p < 0.0001
 Responsivity  149  -8.74  p < 0.0001
 Moderate Intensity Programs  Performance  2,350  -69.32  p < 0.0001
 Responsivity  2,350  -43.00  p < 0.0001
 High Intensity Programs  Performance  389  -34.37  p < 0.0001
 Responsivity  389  -21.38  p < 0.0001
 Community Programs  Performance  17  -2.90  p < 0.01
 Responsivity  17  -2.25  p < 0.05

 Effect Subscale
 Intensity Level  n  M
 Low Intensity Programs  149  0.97
 Moderate Intensity  Programs  2,350  0.90
 High Intensity Programs  389  0.98
 Community Programs  17  1.16

Results for Aboriginal Male Offenders

Similar significant results (p < .05 or greater) emerged for Aboriginal male offenders, with the exception of community programs that had a small sample size and was not significant. These results indicate that Aboriginal offenders are also achieving their expected treatment goals and are performing better at post-program according to facilitator ratings.


Performance and Responsivity Results for Aboriginal Male Offenders
 Performance and Responsivity Subscales
 Intensity Level  Scales  n  t  p
 Low Intensity Programs  Performance  18  -5.84  p < 0.0001
 Responsivity  18  -2.61  p < 0.05
 Moderate Intensity Programs  Performance  438  -32.06  p < 0.0001
 Responsivity  438  -18.75  p < 0.0001
 High Intensity Programs  Performance  147  -20.94  p < 0.0001
 Responsivity  147  -16.47  p < 0.0001
 Community Programs  Performance  13  -5.82  p < 0.0001
 Responsivity  13  -1.49  ns

 Effect Subscale
 Intensity Level  n  M
 Low Intensity Programs  18  0.81
 Moderate Intensity  Programs  438  0.72
 High Intensity Programs  147  0.88
 Community Programs  13  0.52

 

Results for Women Offenders

There was less data available for women offenders than was available for male offenders. There was no data available for low intensity programs or community programs. Despite the small sample sizes, significant difference were found between pre- and post-program indicating that women were also achieving treatment targets at post-program according to facilitator ratings.


Performance and Responsivity Results for Women Offenders
 Performance and Responsivity Subscales
 Intensity Level  Scales  n  t  p
 Moderate Intensity Programs  Performance  9  -5.91  p < 0.001
 Responsivity  9  -2.17  ns
 High Intensity Programs  Performance  80  -13.66  p < 0.0001
 Responsivity  80  -7.43  p < 0.0001

 Effect Subscale
 Intensity Level  n  M
 Moderate Intensity  Programs  9  0.69
 High Intensity Programs  80  0.97

Appendix F: Interview Keys

Correctional Programs Evaluation: Staff Survey

Correctional programs aim to improve offenders reintegration into the community by targeting their criminogenic needs and reducing their level of risk for reoffending. CSC is currently reviewing the continued relevancy, success, and cost-effectiveness of correctional programs, and would like to know about your experiences with correctional programs.

You are being asked to participate in an interview regarding correctional programs that will take approximately 30 minutes. Your participation in the interview is voluntary, and any information you provide will be anonymous. In addition, you are free to refuse to answer any specific questions, and you may withdraw from the study at any time.

Your responses to the interview questions will be anonymous (i.e., your name will not be associated with the interview protocols). Protocols are subject to access as per Section Four (4) of the Access to Information Act. Specifically, no identifying information (e.g., name, position) will be included on the protocols, and your name will not appear in any evaluation report or documentation. Interview responses will be aggregated and presented in summary form within the evaluation report. Therefore, your responses to the survey questions will remain anonymous

Information you provide will offer valuable insight into CSCs correctional programs, as well as help to identify any potential modifications to processes or procedures that might be important for the future.

Do you have any questions?

My signature below indicates that I have read the above, and that I agree to take part in an interview regarding Correctional Programs. The interviewer will also sign to guarantee the conditions stated above.

Date

Participant Name (PRINT)

Participant Signature

***

Date

Interviewer Name (PRINT)

Interviewer Signature

 

Appendix F: SECTION A  IMPLEMENTATION - Page 1

Appendix F: SECTION A  IMPLEMENTATION - Page 2

Appendix F: SECTION A  IMPLEMENTATION - Page 3

Appendix F: SECTION A  IMPLEMENTATION - Page 4

Appendix F: SECTION A  IMPLEMENTATION - Page 5

Appendix F: SECTION A  IMPLEMENTATION - Page 6

Appendix F: SECTION A & B  IMPLEMENTATION & CONTINUED RELEVANCY - Page 7

Appendix F: SECTION B & C  CONTINUED RELEVANCY & QUESTIONS ONLY FOR PROGRAM DELIVERY STAFF- Page 8

Appendix F: SECTION C  QUESTIONS ONLY FOR PROGRAM DELIVERY STAFF - Page 9

Appendix F: SECTION D  BEST PRACTICES & LESSONS LEARNED - Page 10

Appendix F: SECTION D  BEST PRACTICES & LESSONS LEARNED - Page 11

 

Appendix G: Recalibration and Validation of the Spousal Assault Risk Assessment (SARA) as an Actuarial Tool

Introduction

The Spousal Assault Risk Assessment (SARA; Kropp, Hart, Webster, & Eaves, 1994, 1995, 1998) was developed as a set of 20 recommended risk factors to be considered (at a minimum) when evaluating an individuals risk for spousal violence. The 20 risk factors, presented in Table X, were identified on the basis of a review of the relevant scientific literature and a consideration of relevant clinical and legal issues. Part 1 (items 1 through 10) is related to violence risk in general, whereas Part 2 (items 11 through 20) is related specifically to risk of spousal violence. The SARA manual recommends that raters code: (a) the presence of each of the 20 risk factors; (b) the presence of any additional case-specific risk factors; (c) whether any of the risk factors is considered to be critical (that is, particularly relevant to decisions concerning the individuals risk); and (d) the overall degree of risk posed by the individual, taking into account the nature, severity, likelihood, frequency, and imminence of any future violence.

 

Spousal Assault Risk Assessment (SARA) Items
 Part 1 (general violence risk factors)  Part 2 (spousal violence risk factors)
 1. Past assault of family members  11. Past physical assault
 2. Past assault of strangers or acquaintances  12. Past sexual assault/sexual jealousy
 3. Past violation of conditional release or community supervision  13. Past use of weapons and/or credible threats of death
 4. Recent relationship problems  14. Recent escalation in frequency or severity of assault
 5. Recent employment problems  15. Past violation of “no contact” orders
 6. Victim and/or witness to family violence as a child or adolescent  16. Extreme minimization or denial of spousal assault history
 7. Recent substance abuse/dependence  17. Attitudes that support or condone spousal assault
 8. Recent suicidal or homicidal ideation/intent  18. Severe and/or sexual assault (most recent incident)
 9. Recent psychotic and/or manic symptoms  19. Use of weapons and/or credible threats of death (most recent incident)
 10. Personality disorder with anger, impulsivity, or behavioural instability  20. Violation of –no contact– order (most recent incident)

The endorsement of individual SARA risk factors is coded numerically, and item codes can be combined arithmetically to yield several continuous scores (described below) that may be of use in making descriptive judgments or in research. Some argue that, because the scores are combined according to fixed and explicit algorithms, they can be considered actuarial in nature (Kropp & Hart, 2000). However, the summary ratings of risk derived from these item scores are not made according to fixed and explicit guidelines. As such, the summative results of the SARA are based largely on what is referred to as structured professional judgment.

 

Purpose of the Present Recalibration

The purpose of the present exercise was to evaluate whether the efficiency and predictive accuracy of the SARA can be improved with the development of a proxy version and recalibrated version of the SARA, based on automated offender data. By adjusting the scoring of the SARA in such a way as to provide an overall actuarial assessment of spousal assault risk, it was hypothesized that the recalibrated version of the SARA would provide a more efficient, reliable and valid assessment of risk for spousal assault.

 

Method

Sample Composition

For the purposes of this study, all available data for federally sentenced non-Aboriginal male offenders who had a SARA rating and did not participate in CSCs FVPP were extracted from CSCs automated database (OMS). As of March 2008, information was available for N = 756 non-Aboriginal male offenders released from federal institutions between 2000 and 2007 and available for a follow up period of 3 years.

 

Measures

Spousal Assault Risk Assessment (SARA)

SARA ratings are based on interviews with offenders and a review of all relevant file information. The presence of the 20 risk factors is scored on a 3-point scale (0 = absent, 1 = possibly or partially present, 2 = present). When items are omitted due to missing information, they are coded as absent. Critical items ratings are made on a 2–point scale (0 = not critical, 1 = critical). Three continuous scores are calculated from the item ratings: (a) Total Scores reflect the sum of individual item ratings, ranging from 0 to 40; (b) Number of Factors Present is an index of how many risk factors are rated as present (i.e., 2 vs. 0 or 1) in a given case, ranging from 0 to 20; and (c) Number of Critical Items is an index of how many risk factors were rated critical, ranging from 0 to 20. Finally, raters code Summary Risk Ratings, reflecting their judgments concerning offenders risk for recidivistic spousal assault on a 3–point scale (0 = low risk, 1 = moderate risk, 2 = high risk).

The Recalibrated SARA

The recalibrated SARA was derived by randomly dividing the sample into two equal groups; the first sub–sample (N = 378) was used for the purpose of re–calibration. The Burgess method was used to re–weight the SARA–Proxy scale items on this sample. Next, the re–calibrated SARA was validated using the second equal–sized sub-sample. Cronbachs alpha reliability coefficient, Pearson correlation coefficients, and Receiver Operating Characteristics (ROC) statistics were used to measure the reliability and predictive validity of the recalibrated SARA.

 

Procedures

SARA

SARA scores for the release cohort of federally sentenced non–Aboriginal male offender population were obtained from CSCs automated OMS. Offender identifiers for this group were matched to those in OMS data relations containing SARA information.

The Recalibrated SARA

The male offender release cohort was randomly divided into two equally–sized groups. Next, the Burgess method was used to recalibrate the SARA–Proxy items. This method scores endorsed items based on differences between the overall sample recidivism rate and that of those endorsing a particular item. The scoring technique assigns a 1 for every 5% between the overall and item–associated rates. This scoring system begins with differences greater than or less than the overall sample mean plus/minus 5%. For example, the recidivism rate (return with a new offence) of the random sample was 28.89%. Offenders who endorsed item 7 of the SARA (Recent Substance Abuse/Dependence) as “Serious substance abuse or substance dependence within the past year” had a recidivism rate of 15.38%. Thus, this sub–item of the recalibrated SARA was scored as follows:

Item 7 Serious substance abuse Score = ((28.89–5) 15.38)/5 = 1.7; rounded = 2

Next, risk groupings were established by ranking the recalibrated scores into five equal clusters. Corresponding cut–off scores were created based on these groupings.

Finally, the recalibrated item scores and risk groupings were applied to the second equal–sized sample.

 

Results

The Recalibrated SARA (SARA–R): Validation

Internal consistency of the SARA–R was tested using Cronbachs alpha reliability coefficient. This yielded raw alphas of 0.32. The cut–off scores for the SARA–proxy groupings were established to reflect the sample distribution.

When validated on outcome as a measure of predictive validity, the SARA–R had an AUC of 0.56 (p < .05).

There were four individual items for which there was no variation from the overall recidivism rate for all classes within the item, thus no scores were associated with that item using the Burgess method. These items were:




 

Appendix H: Pre-/Post-Program Assessment Battery Results

Pre– Post Limitations

The pre– and post–program assessment batteries analysed in this evaluation originally came from either Reintegration Programs or the Research Branch. The data is collected in the field (i.e., institutions and community offices where the programs are delivered) and either entered into spreadsheets on–site or sent to NHQ to be entered manually. All the data, including site location, program start and end dates, and Finger Print System (FPS) number were all entered manually and not linked to OMS. This can cause data quality issues where FPS numbers are not entered accurately and the data cannot be connected to the correct offender, data are missing, values are out of range, pre– and post– cannot clearly be deciphered, and other complications. This resulted in the omission of some data.

In all correctional programs, there was a much larger sample that completed the program but for which we do not have their pre– and post–program data. It is unclear if this is because the data were not sent to NHQ or because the offenders did not complete the data. However, given the smaller sample of offenders for which there is pre–/post-program data available, the generalizability of the sample to the larger group may be in question.

A technique was employed to substitute for missing data. This technique, called multiple imputation, is a well–validated method of dealing with missing data. Imputation is essentially a statistical technique used to examine all data in the data set and assign the most likely data point for the missing data. A cut–off point of 20% was established where, if the offender had 20% or less missing data, imputation would occur. However, if the offender had 21% or more missing data, their subscale total was eliminated from analysis. This technique was implemented for all pre– and post–program assessment batteries except the NSAP data that was not imputed. FVPP datasets employed a similar data imputation technique using the mean rather than a statistical algorithm to estimate the missing values.

 

Violent Offender Programs

VPP Pre–/Post Assessment Battery Results

The assessment battery for VPP contains four instruments: (a) Aggression Questionnaire; (b) Balanced Inventory of Desirable Responding; (c) Impulsivity Scale; and (d) Reactions to Provocations.

Aggression Questionnaire (AQ; Buss & Parry, 1992)

The AQ is a 29–item self-report measure completed pre– and post–program assessing offenders attitudes and feelings around anger, hostility and aggression. The instrument has four subscales: (a) physical aggression (involves hurting or harming others); (b) verbal aggression (involves harming others); (c) anger (involves physiological arousal and preparation for aggression; and (d) hostility (feelings of ill will and injustice).

Reliability and Validity

No reliability and validity information is available on this instrument.

Results for non-Aboriginal Male Offenders

Results for the sample of 163 non–Aboriginal male offenders who completed pre–/post-test assessment indicated significant pre-post program results. Results on all four subscales were very significant. This suggests that offenders, through participation in the program, changed some attitudes and beliefs regarding aggression.


 Subscale  t value  N value  p value
 Physical aggression  10.70  163  p < 0.0001
 Verbal aggression  8.86  163  p < 0.0001
 Anger  9.96  163  p < 0.0001
 Hostility  10.84  163  p < 0.0001

Results for Aboriginal Male Offenders

Results for the sample of 65 Aboriginal male offenders who completed pre-/post test assessment indicated significant pre-/post-program results. Results on all four subscales were very significant. This suggests that offenders, through participation in the program, changed some attitudes and beliefs regarding aggression.

 Subscale  t value  N value  p value
 Physical aggression  7.54  65  p < 0.0001
 Verbal aggression  3.32  65  p < 0.001
 Anger  5.02  65  p < 0.0001
 Hostility  4.66  65  p < 0.0001

 

Balanced Inventory of Desirable Responding (BIDR; Paulhus, 1984)

The BIDR is 40–item self-report measure that measures two constructs: (a) self–deception (the tendency to give self–reports that are honest but positively biased); and (b) impression management (deliberate self-presentation to an audience). The impression management scale was developed under the assumption that some respondents systematically under report their undesirable behaviours, and in extreme cases, respondents are considered to be “faking good” (i.e. final scores 12 and over). Such styles impact the confidence that can be placed on self–report measures, but this does not mean they should be disregarded.

Reliability and Validity

The internal reliability or the extent to which the scale consistently measures the same construct is highly satisfactory. Both subscales have been found to have good convergent validity.

Results for non–Aboriginal Male Offenders

Given that high scores on the impression management domain may be more suspect, it was important to examine the proportion of offender who scored 12 or higher (i.e. may be ‘faking good’). For the impression management scale, pre– and post-program values indicate only a small percentage of offenders may have been faking good (3.4% (n = 7) at pre–program and 7.7% (n = 16) at post-program.

Results for Aboriginal Male Offenders

Similar rates emerged for Aboriginal offenders. For the impression management scale, pre– and post–program values indicate only a small percentage of offenders may have been ‘faking good’ (3.8% (n = 3) at pre–program and 10.1% (n = 8) at post–program.

 

Impulsivity Scale (IS; Eyesenck, 1985)

The IS is a 19–item self-report measure that measures impulsivity. There are no subscales wherein the total score is used for interpretation. Higher scores indicate delays in impulsive violence and promotion of long-term planning.

Reliability and Validity

No reliability and validity information is available for this instrument.

Results for non–Aboriginal Male Offenders

Results indicate that for the 199 non–Aboriginal offenders who completed the measure pre– and post–program, the test was highly significant (p < .0001). This result seems to indicate that offenders were able to reduce their impulsive thoughts.

Results for Aboriginal Male Offenders

Results indicate that for the 77 Aboriginal offenders who completed the measure pre– and post-program, the test was highly significant (p < .0001).

 

Reactions to Provocations (RP; Novaco, 1994)

The RP is a self–report measure that contains two parts. Part A is a 48–item questionnaire that provides statements describing things that people think, feel, and do. Part A has 12 subscales: (a) attentional focus; (b) rumination; (c) hostile attitude; (d) suspicion; (e) intensity; (f) duration; (g) somatic activation/tension; (h) irritability; (i) impulsive reaction; (j) verbal aggression; (k) physical confrontation; and (l) indirect expression. Part B of the test has 25 questions and examines the amount of anger the participant would feel if the situation were to happen to him/her. Part B has 5 subscales: (a) disrespectful treatment; (b) unfairness/injustice; (c) frustration/interruption; (d) annoying traits; and (e) irritations.

Reliability and Validity

Reliability was indicated through high test–retest correlations (exceeding.80; Mills, Kroner & Forth, 1998). The subscales demonstrated a stronger relationship with their respective domains than with other domains indicating construct validity. There was also a strong relationship found between this scale and other anger and aggression measures, providing evidence for concurrent validity.

Results for Non–Aboriginal Male Offenders

Results indicate that on all 17 subscales, there were significant differences between pre– and post-program, indicating that offenders changed their thinking around responding to provocation.

 Subscale    t value  N value  p value
 Part
 A  Attitudinal focus  7.50  198  p < 0.0001
 A  Rumination  9.79  197  p < 0.0001
 A  Hostile attitude  10.97  199  p < 0.0001
 A  Suspicion  7.36  196  p < 0.0001
 A  Intensity  9.65  197  p < 0.0001
 A  Duration  8.12  199  p < 0.0001
 A  Somatic activation/ tension  8.06  196  p < 0.0001
 A  Irritability  8.47  198  p < 0.0001
 A  Impulsive reaction  9.78  198  p < 0.0001
 A  Verbal aggression  10.81  199  p < 0.0001
 A  Physical confrontation  9.96  194  p < 0.0001
 A  Indirect expression  9.84  198  p < 0.0001
 B  Disrespectful treatment  9.45  200  p < 0.0001
 B  Unfairness/injustice  9.03  200  p < 0.0001
 B  Frustration/interruption  9.48  200  p < 0.0001
 B  Annoying traits  9.37  197  p < 0.0001
 B  Irritations  8.17  199  p < 0.0001

Results for Aboriginal Male Offenders

Very similar results emerge for Aboriginal offenders wherein on all 17 subscales, there were significant differences between pre- and post-program, indicating that offenders changed their thinking around responding to provocation.

 Subscale    t value  N value  p value
 Part
 A  Attitudinal focus  2.36  75  p < 0.05
 A  Rumination  4.19  74  p < 0.0001
 A  Hostile attitude  4.90  75  p < 0.0001
 A  Suspicion  4.26  75  p < 0.0001
 A  Intensity  5.12  74  p < 0.0001
 A  Duration  5.60  75  p < 0.0001
 A  Somatic activation/ tension  2.41  75  p < 0.05
 A  Irritability  3.74  74  p < 0.0001
 A  Impulsive reaction  5.85  75  p < 0.0001
 A  Verbal aggression  4.96  74  p < 0.0001
 A  Physical confrontation  5.51  74  p < 0.0001
 A  Indirect expression  4.04  74  p < 0.0001
 B  Disrespectful treatment  5.21  76  p < 0.0001
 B  Unfairness/injustice  4.49  76  p < 0.0001
 B  Frustration/interruption  5.45  76  p < 0.0001
 B  Annoying traits  3.90  75  p < 0.0001
 B  Irritations  4.26  76  p < 0.0001

Family Violence Prevention Programs

High Intensity Family Violence Pre-/Post Assessment Battery Results

The Assessment battery for the HIFVPP has been modified somewhat since the beginning of the program. The assessment batteries that will be examined here are part of the current assessment manual and have a pre- and a post-program assessment. As such, five instruments will be examined: (a) Domestic Violence Vignettes; (b) Goal Attainment Scale Family Violence; (c) Interpersonal Relationships Scale; (d) Abusive Relationships Inventory; and (e) Paulhus Deception Scale.

 

Domestic Violence Vignettes (DVV; Stewart & Gabora, 2006)

These data are collected by the program facilitators in a structured interview. The questions evaluate how the offender would respond to a number of potential risk situations related to family violence. Vignettes are administered pre- and post-program to measure situations that involve: (a) jealousy; (b) control issues; (c) rejection; (d) sexual issues; and (e) employment and finances. There is also a total subscale. In addition to the five subscales, skills are assessed in the following three domains: (a) behavioural response; (b) interpretation; and (c) perspective taking. It should be noted that the Perspective-Taking subscale was missing from the database and as a result could not be analyzed. Scores are expected to increase indicating that the respondent has good recognition and good application of the required scores.

Reliability and Validity

The vignettes are clinical assessment tools, which have been shown to predict recidivism and correlate with belief scales in other applications.

Results for Non-Aboriginal Offenders

Results reveal that offenders had significant results on all subscales of the DVV (p < .0001) with the exception of the jealousy index that was not significant. These results seem to indicate that non-Aboriginal respondents have learned good recognition and good application of skills learned in the program through participation.

 Subscale  t value  N value  p value
 Jealousy  -1.32  172  ns
 Rejection  -10.33  165  p < 0.0001
 Sexual issues  -5.95  165  p < 0.0001
 Employment and finances  -8.64  165  p < 0.0001
 Control issues  -7.05  165  p < 0.0001
 Behavioural response  -11.45  163  p < 0.0001
 Interpretation  -10.97  163  p < 0.0001
 Total  -10.21  165  p < 0.0001

Results for Aboriginal Offenders

Results reveal that offenders had significant results on all subscales of the DVV (p < .0001) with the exception of the jealousy index that was not significant. These results seem to indicate that Aboriginal respondents have learned good recognition and good application of skills learned in the program through participation.

 Subscale  t value  N value  p value
 Jealousy  -0.60  86  ns
 Rejection  -6.60  84  p < 0.0001
 Sexual issues  -5.16  84  p < 0.0001
 Employment and finances  -5.53  84  p < 0.0001
 Control issues  -6.99  84  p < 0.0001
 Behavioural response  -9.80  80  p < 0.0001
 Interpretation  -7.96  80  p < 0.0001
 Total  -8.95  84  p < 0.0001

 

Goal Attainment Scale Family Violence (GAS-FV; Stewart & Gabora, 1999)

The GAS–FV assesses the treatment goals of the program participants. This instrument is completed by the program facilitators at pre–, mid–, and post-program. Each of the 14 questions assesses different program targets such as problematic thinking, underlying relationship violence, emotional mismanagement, problems with self-regulation, and deficits in social communication. The analyses reported here examine pre–/post–program differences.

Reliability and Validity

No information is available.

Results for Non-Aboriginal Offenders

All subscale results are significant (p < .0001) indicating that non-Aboriginal offenders did move toward attaining program targets such as understanding problematic thinking, underlying relationship violence, emotional mismanagement, problems with self–regulation, and deficits in social communication.

 Subscale  t value  N value  p value
 Acceptance of responsibility for abusive/ violent  behaviour  -16.75  175  p < 0.0001
 Acknowledges use of power and control tactics  -20.04  175  p < 0.0001
 Shows empathy for victim  -13.83  175  p < 0.0001
 Extent of skills development  -16.74  175  p < 0.0001
 Recognises cognitive distortions beliefs and  neutralizations  -21.33  175  p < 0.0001
 Minimises consequences  -16.75  175  p < 0.0001
 Understands lifestyle dynamics  -19.02  175  p < 0.0001
 Understands abusive patterns  -20.82  173  p < 0.0001
 Identify relapse prevention concepts  -21.56  164  p < 0.0001
 Discloses personal information  -12.29  173  p < 0.0001
 Participation in treatment  -10.41  171  p < 0.0001
 Motivation to change behaviour  -11.23  175  p < 0.0001
 Overall quality of plan on release  -16.74  160  p < 0.0001
 Overall participation and program performance  -13.65  160  p < 0.0001

Results for Aboriginal Offenders

Results were the same for Aboriginal offenders also indicating that the offenders moved toward attaining program targets such as understanding problematic thinking, underlying relationship violence, emotional mismanagement, problems with self–regulation, and deficits in social communication.

 Subscale  t value  N value  p value
 Acceptance of responsibility for abusive/ violent  behaviour  -14.60  83  p < 0.0001
 Acknowledges use of power and control tactics  -13.57  83  p < 0.0001
 Shows empathy for victim  -13.16  83  p < 0.0001
 Extent of skills development  -12.28  83  p < 0.0001
 Recognises cognitive distortions beliefs and  neutralizations  -15.80  83  p < 0.0001
 Minimises consequences  -12.36  83  p < 0.0001
 Understands lifestyle dynamics  -11.22  83  p < 0.0001
 Understands abusive patterns  -13.18  83  p < 0.0001
 Identify relapse prevention concepts  -16.15  83  p < 0.0001
 Discloses personal information  -11.66  83  p < 0.0001
 Participation in treatment  -7.87  83  p < 0.0001
 Motivation to change behaviour  -8.19  83  p < 0.0001
 Overall quality of plan on release  -13.71  81  p < 0.0001
 Overall participation and program performance  -9.51  81  p < 0.0001

 

Interpersonal Relationship Scale (IRS; Hupka & Rusch, 2001)

The IRS is a self-report measure containing six subscales, each examining a different aspect of jealousy. The six subscales are: (a) threat to exclusive companionship; (b) self-deprecation/envy; (c) dependency; (d) sexual possessiveness; (e) competition and vindictiveness; and (f) distrust. Results are expected to increase with program participation.

Reliability and Validity

No information is available.

Results for Non-Aboriginal Offenders

Results indicate for non-Aboriginal offenders that all subscales of the IRS were significant (p < .0001), indicating that non-Aboriginal offenders are reporting reduced levels of jealousy.

 Subscale  t value  N value  p value
 Threat to exclusive  companionship  -6.96  177  p < 0.0001
 Self–deprecation/envy  -5.97  174  p < 0.0001
 Dependency  -8.41  180  p < 0.0001
 Sexual possessiveness  -5.45  179  p < 0.0001
 Competition and vindictiveness  -8.36  169  p < 0.0001
 Distrust  -4.77  186  p < 0.0001

Results for Aboriginal Offenders

Results for Aboriginal offenders indicate that 5 of the 6 subscales of the IRS were significant (p   .05 and better) indicating that Aboriginal offenders also reported reduced levels of jealousy. The one subscale that was not significantly different for Aboriginal offenders was distrust, possibly indicating that this construct did not change as a result of program participation.

 Subscale  t value  N value  p value
 Threat to exclusive  companionship  -4.33  85  p < 0.0001
 Self–deprecation/envy  -3.80  85  p < 0.001
 Dependency  -4.95  85  p < 0.0001
 Sexual possessiveness  -2.20  87  p < 0.05
 Competition and vindictiveness  -4.85  82  p < 0.0001
 Distrust  -1.50  86  ns

 

Abusive Relationship Inventory (ARI; Boer, Kroner, Wong & Cadsky, n.d)

The ARI is a self-report measure containing four subscales: (a) rationales for hitting; (b) need for control; (c) legal entitlement; and (d) batterers myths. The subscales measure attitudes supportive of abuse in intimate relationships. Scores are expected to decrease with program participation.

Reliability and Validity

The ARI has demonstrated good reliability on all four subscales (Boer, Kroner, Wong & Cadsky, n.d.).

Results for Non-Aboriginal Offenders

Results from the ABI indicated significant results (p < .0001) for non-Aboriginal offenders on all four subscales. These results suggest that program participation may reduce attitudes and beliefs associated with the support of intimate violence.

 Subscale  t value  N value  p value
 Rationales for hitting  4.98  181  p < 0.0001
 Need for control  10.10  181  p < 0.0001
 Legal entitlement  6.52  179  p < 0.0001
 Batterers myths  6.70  181  p < 0.0001

Results for Aboriginal Offenders

The same results were found for Aboriginal offenders on all four subscales of the ARI. These results suggest that program participation may reduce attitudes and beliefs associated with the support of intimate violence.

 Subscale  t value  N value  p value>
 Rationales for hitting  3.62  85  p < 0.001
 Need for control  5.04  85  p < 0.0001
 Legal entitlement  2.14  84  p < 0.05
 Batterers myths  4.96  84  p < 0.0001

 

Paulhus Deception Scale (PDS; Paulhus, 1990)

The PDS is 40–item self–report measure that measures two constructs: (a) self–deception (the tendency to give self–reports that are honest but positively biased); and (b) impression management (deliberate self–presentation to an audience). The impression management scale was developed under the assumption that some respondents systematically under report their undesirable behaviours, and in extreme cases, respondents are considered to be “faking good” (i.e., final scores 12 and over). Such styles impact the confidence that can be placed on self-report measures, but this does not mean they should be disregarded.

Reliability and Validity

The internal reliability or the extent to which the scale consistently measures the same construct is highly satisfactory. Both subscales have been found to have good convergent validity.

Results for non–Aboriginal Male Offenders

Given that high scores on the impression management domain may be more suspect, it was important to examine the proportion of offender who scored 12 or higher (i.e., may be faking good). For the impression management scale, pre– and post–program values indicated that only a small percentage of offenders may have been faking good (12.3%; N = 44) at pre–program.

Results for Aboriginal Male Offenders

Similar rates emerged for Aboriginal offenders. For the impression management scale, pre– and post–program values indicate only a small percentage of offenders may have been faking good (10.9%; N = 11) at pre–program.

Moderate Intensity Family Violence Pre- Post Assessment Battery Results

The assessment battery for the MIFVPP has been modified somewhat since the beginning of the program. The assessment batteries that are examined here are part of the current assessment manual and have a pre– and a post-program assessment. As such, five instruments were examined: (a) Domestic Violence Vignettes; (b) Goal Attainment Scale Family Violence; (c) Interpersonal Relationships Scale; (d) Abusive Relationships Inventory; and (e) Paulhus Deception Scale.

 

Domestic Violence Vignettes (DVV; Stewart & Gabora, 2006)

These data are collected by the program facilitators in a structured interview. The questions evaluate how the offender would respond to a number of potential risk situations related to family violence. Vignettes are administered pre– and post–program to measure situations that involve: (a) jealousy; (b) control issues; (c) rejection; (d) sexual issues; and (e) employment and finances. There is also a total subscale. In addition to the five subscales, skills are assessed in the following three domains: (a) behavioural response; (b) interpretation; (c) perspective taking. It should be noted that the Perspective–Taking subscale was missing from the database and as a result could not be analyzed. Scores are expected to increase indicating that the respondent has good recognition and good application of the required scores.

Reliability and Validity

The vignettes are clinical assessment tools, which have been shown to predict recidivism and correlate with belief scales in other applications.

Results for Non–Aboriginal Offenders

Results reveal that offenders had significant results on all subscales of the DVV (p < .0001). These results seem to indicate that non–Aboriginal respondents have learned good recognition and application of skills learned through program participation.

 Subscale  t value  N value  p value
 Jealousy  -11.20  463  p < 0.0001
 Rejection  -16.88  463  p < 0.0001
 Sexual issues  -16.96  463  p < 0.0001
 Employment and finances  -16.23  463  p < 0.0001
 Control issues  -15.15  463  p < 0.0001
 Behavioural response  -22.22  430  p < 0.0001
 Interpretation  -20.17  400  p < 0.0001
 Total  -22.88  463  p < 0.0001

Results for Aboriginal Offenders

Results reveal that offenders had significant results on all subscales of the DVV (p < .0001). These results seem to indicate that Aboriginal respondents have learned good recognition and application of skills learned through program participation.

 Subscale  t value  N value  p value
 Jealousy  -9.50  133  p < 0.0001
 Rejection  -8.04  133  p < 0.0001
 Sexual issues  -6.64  133  p < 0.0001
 Employment and finances  -7.47  133  p < 0.0001
 Control issues  -6.54  133  p < 0.0001
 Behavioural response  -11.77  125  p < 0.0001
 Interpretation  -10.23  113  p < 0.0001
 Total  -11.57  133  p < 0.0001

 

Goal Attainment Scale Family Violence (GAS-FV; Stewart & Gabora, 1999)

The GAS–FV assesses the treatment goals of the program participants. This instrument is completed by the program facilitators at pre–, mid–, and post–program. Each of the 14 questions assesses different program targets such as problematic thinking, underlying relationship violence, emotional mismanagement, problems with self–regulation, and deficits in social communication. The analyses reported here examine pre–/post-program differences.

Reliability and Validity

No information is available.

Results for Non–Aboriginal Offenders

All GAS–FV subscale results are significant (p < .0001) indicating that non-Aboriginal offenders did move toward attaining program targets such as understanding and shifting attitudes around problematic thinking, underlying relationship violence, and emotional mismanagement, problems with self–regulation, and deficits in social communication.

 Subscale  t value  N value  p value
 Acceptance of responsibility for abusive/ violent  behaviour  -23.98  481  p < 0.0001
 Acknowledges use of power and control tactics  -29.31  481  p < 0.0001
 Shows empathy for victim  -23.71  481  p < 0.0001
 Extent of skills development  -23.13  479  p < 0.0001
 Recognises cognitive distortions beliefs and  neutralizations  -31.12  470  p < 0.0001
 Minimises consequences  -25.10  481  p < 0.0001
 Understands lifestyle dynamics  -27.73  479  p < 0.0001
 Understands abusive patterns  -30.28  475  p < 0.0001
 Identify relapse prevention concepts  -30.74  468  p < 0.0001
 Discloses personal information  -21.92  481  p < 0.0001
 Participation in treatment  -18.51  481  p < 0.0001
 Motivation to change behaviour  -21.32  481  p < 0.0001
 Overall quality of plan on release  -26.24  447  p < 0.0001
 Overall participation and program performance  -20.19  468  p < 0.0001

Results for Aboriginal Offenders

Results were the same for Aboriginal offenders which also indicated that the offenders moved toward attaining program targets such as understanding and shifting attitudes around problematic thinking, underlying relationship violence, emotional mismanagement, problems with self–regulation, and deficits in social communication.

 Subscale  t value  N value  p value
 Acceptance of responsibility for abusive/ violent  behaviour  -12.08  137  p < 0.0001
 Acknowledges use of power and control tactics  -14.56  137  p < 0.0001
 Shows empathy for victim  -10.82  137  p < 0.0001
 Extent of skills development  -10.36  137  p < 0.0001
 Recognises cognitive distortions beliefs and  neutralizations  -14.62  137  p < 0.0001
 Minimises consequences  -11.09  137  p < 0.0001
 Understands lifestyle dynamics  -12.98  137  p < 0.0001
 Understands abusive patterns  -16.59  137  p < 0.0001
 Identify relapse prevention concepts  -14.55  137  p < 0.0001
 Discloses personal information  -11.09  137  p < 0.0001
 Participation in treatment  -9.36  137  p < 0.0001
 Motivation to change behaviour  -10.27  137  p < 0.0001
 Overall quality of plan on release  -11.79  135  p < 0.0001
 Overall participation and program performance  -10.18  137  p < 0.0001

 

Interpersonal Relationship Scale (IRS; Hupka & Rusch, 2001)

The IRS is a self–report measure containing six subscales, each examining a different aspect of jealousy. The subscales are: (a) threat to exclusive companionship; (b) self–deprecation/envy; (c) dependency; (d) sexual possessiveness; (e) competition and vindictiveness; and (f) distrust. Results are expected to increase with program participation.

Reliability and Validity

No information is available.

Results for Non-Aboriginal Offenders

Results indicate that all subscales of the IRS were significant (p < .0001) indicating that non-Aboriginal offenders are reporting reduced levels of jealousy.

 Subscale  t value  N value  p value
 Threat to exclusive  companionship  -8.33  481  p < 0.0001
 Self–deprecation/envy  -8.16  478  p < 0.0001
 Dependency  -10.81  480  p < 0.0001
 Sexual possessiveness  -5.87  481  p < 0.0001
 Competition and vindictiveness  -8.98  478  p < 0.0001
 Distrust  -5.11  479  p < 0.0001

Results for Aboriginal Offenders

Results indicate that all subscales of the IRS were significant (p < .01 and better) indicating that Aboriginal offenders are also reporting reduced levels of jealousy.

 Subscale  t value  N value  p value
 Menace à l’exclusivité de la relation  -4,61  137  p < 0,0001
 Autodépréciation/envie  -3,98  134  p < 0,0001
 Dépendance  -4,73  136  p < 0,0001
 Possessivité sexuelle  -3,18  137  p < .01
 Compétition et caractère vindicatif  -5,41  133  p < 0,0001
 Méfiance  -3,93  137  p < 0,0001

 

Abusive Relationship Inventory (ARI; Boer, Kroner, Wong & Cadsky, n.d.)

The ARI is a self-report measure containing four subscales: (a) rationales for hitting; (b) need for control; (c) legal entitlement; and (d) batterers myths. The subscales measure attitudes supportive of abuse in intimate relationships. Scores are expected to decrease with program participation.

Reliability and Validity

The ARI has demonstrated good reliability on all four subscales (Boer, Kroner, Wong & Cadsky, n.d.).

Results for Non-Aboriginal Offenders

Results from the ARI indicated significant results (p < .0001) for non-Aboriginal offenders on all four subscales. These results suggest that program participation may reduce attitudes and beliefs associated with the support of intimate violence.

 Subscale  t value  N value  p value
 Rationales for hitting  5.01  491  p < 0.0001
 Need for control  10.82  491  p < 0.0001
 Legal entitlement  4.31  491  p < 0.05
 Batterers myths  7.58  490  p < 0.0001

Results for Aboriginal Offenders

Results for Aboriginal offenders on the ARI indicated significant results (p < .001) on three of the four subscales. These results suggest that program participation may reduce attitudes and beliefs associated with the support of intimate violence. There were no significant results for legal entitlement possibly suggesting that Aboriginal offenders did not change their attitudes and beliefs around the legal entitlement of intimate partner violence.

 Subscale  t value  N value  p value
 Rationales for hitting  3.28  140  p < 0.001
 Need for control  5.74  140  p < 0.001
 Legal entitlement  1.39  139  ns
 Batterers myths  3.44  140  p < 0.001

 

Paulhus Deception Scale (PDS; Paulhus, 1990)

The PDS is 40–item self–report measure that measures two constructs: (a) self–deception (the tendency to give self–reports that are honest but positively biased); and (b) impression management (deliberate self–presentation to an audience). The impression management scale was developed under the assumption that some respondents systematically under report their undesirable behaviours, and in extreme cases, respondents are considered to be “faking good” (i.e. final scores 12 and over). Such styles impact the confidence that can be placed on self–report measures, but this does not mean they should be disregarded.

Reliability and Validity

The internal reliability or the extent to which the scale consistently measures the same construct is highly satisfactory. Both subscales have been found to have good convergent validity.

Results for non–Aboriginal Male Offenders

Given that high scores on the impression management domain may be more suspect, it was important to examine the proportion of offender who scored 12 or higher (i.e. may be ‘faking good’). For the impression management scale, pre– and post–program values indicate only a small percentage of offenders may have been ‘faking good’ (11.6% (n = 86) at pre–program.

Results for Aboriginal Male Offenders

Similar rates emerged for Aboriginal offenders. For the impression management scale, pre– and post–program values indicate only a small percentage of offenders may have been ‘faking good’ (9.2% (n = 14) at pre–program.

National Substance Abuse Pre–/Post–Program Assessment Battery

The NSAP program has 8 scales selected to inform on the offender progress toward treatment targets. The assessments measures are: (a) Inventory of Drug Taking Situations; (b) Drug Taking Confidence Questionnaire; (c) Effectiveness of Coping Behaviours Inventory; (d) Drug Avoidance Self Efficacy Scale; (e) Craving Beliefs Questionnaire; (f) Beliefs About Substance Abuse; (g) Drinking and Drug Related Locus of Control; and (h) Paulhus Deception Scales.

 

Inventory of Drug Taking Situations (IDTS; Annis, Turner & Sklar, 1997)

The IDTS produces a profile of a respondents substance use over the last 12–months and a profile of risk taking situations. Items load onto two categories of drug use situations: (a) personal states; and (b) situations involving other people. Under the personal states category there are 5 subscales: (a) unpleasant emotions; (b) physical discomfort; (c) pleasant emotions; (d) testing personal control; and (e) urges/temptations. Under the situations involving other people category, there are three subscales: (a) conflict with others; (b) social pressure to use; and (c) pleasant times with others. It consists of two questionnaires, one for alcohol and one for up to three types of drugs.

Results for Non–Aboriginal Male Offenders

Results for Non–Aboriginal Offenders on Inventory of Drug Taking Situations
 Subscale  t value  N value  p value
 Unpleasant emotions  2.53  161  p < 0.05
 Physical discomfort  0.81  161  ns
 Pleasant emotions  2.30  161  p < 0.05
 Testing personal control  1.19  161  ns
 Urges/temptations  1.73  161  p < 0.10
 Conflict with others  2.53  161  p < 0.05
 Social pressure to use  0.85  161  ns
 Pleasant times with others  0.85  161  ns

 

Results for Non–Aboriginal Offenders on Inventory of Drug Taking Situations
 Subscale  t value  N value  p value
 Unpleasant emotions  0.16  201  ns
 Physical discomfort  –1.35  201  ns
 Pleasant emotions  –0.19  201  ns
 Testing personal control  –2.38  201  p < 0.01
 Urges/temptations  –0.22  201  ns
 Conflict with others  0.16  201  ns
 Social pressure to use  –1.27  201  ns
 Pleasant times with others  –1,27  201  ns

 

IDTS, Non–Aboriginal Male, Alcohol, Moderate
 Subscale  t value  N value  p value
 Unpleasant emotions  5.22  635  p < 0.0001
 Physical discomfort  2.37  635  p < 0.05
 Pleasant emotions  5.03  635  p < 0.0001
 Testing personal control  1.44  635  ns
 Urges/temptations  3.81  635  p < 0.001
 Conflict with others  5.22  635  p < 0.0001
 Social pressure to use  4.34  634  p < 0.0001
 Pleasant times with others  4.34  634  p < 0.0001

 

IDTS, Non–Aboriginal Male, Drugs, Moderate">
 Subscale  t value  N value  p value
 Unpleasant emotions 10.38  1,279  p < 0.0001
 Physical discomfort  7.78  1,279  p < 0.0001
 Pleasant emotions  6.22  1,278  p < 0.0001
 Testing personal control  2.60  1,279  p < 0.001
 Urges/temptations  8.08  1,279  p < 0.0001
 Conflict with others  10.38  1,279  p < 0.0001
 Social pressure to use  6.18  1,279  p < 0.0001
 Pleasant times with others  6.18  1,279  p < 0.0001

 

Results for Non–Aboriginal Offenders on Inventory of Drug Taking Situations
 Subscale  t value  N value  p value
 Unpleasant emotions 2.94  138  p < 0.01
 Physical discomfort  0.33  138  ns
 Pleasant emotions  1.94  138  p < 0.10
 Testing personal control  –0.32  138  ns
 Urges/temptations  2.24  138  p < 0.05
 Conflict with others  2.94  138  p < 0.01
 Social pressure to use  3.17  138  p < 0.01
 Pleasant times with others  3.17  138  p < 0.01

 

Results for Non–Aboriginal Offenders on Inventory of Drug Taking Situations
 Subscale  t value  N value  p value
 Unpleasant emotions 6.28  321  p < 0.0001
 Physical discomfort  3.90  321  p < 0.0001
 Pleasant emotions  3.36  321  p < 0.001
 Testing personal control  1.72  321  p < 0.10
 Urges/temptations  5.40  321  p < 0.0001
 Conflict with others  6.28  321  p < 0.0001
 Social pressure to use  4.51  321  p < 0.0001
 Pleasant times with others  4.51  321  p < 0.0001

Results for Aboriginal Male Offenders

IDTS, Aboriginal Male, Alcohol, Low
 Subscale  t value  N value  p value
 Unpleasant emotions 0.98  27  ns
 ÉPhysical discomfort  2.50  27  p < 0.05
 Pleasant emotions  2.20  27  p < 0.05
 Testing personal control  1.58  27  ns
 Urges/temptations  0.71  27  ns
 Conflict with others  0.98  27  ns
 Social pressure to use  1.19  27  ns
 Pleasant times with others  1.19  27  ns

 

IDTS, Aboriginal Male, Drugs, Low
 Subscale  t value  N value  p value
 Unpleasant emotions 1.57  25  ns
 Physical discomfort  1.07  25  ns
 Pleasant emotions  1.15  25  ns
 Testing personal control  3.00  25  p < 0.01
 Urges/temptations  1.37  25  ns
 Conflict with others  1.57  25  ns
 Social pressure to use  0.67  25  ns
 Pleasant times with others  0.67  25  ns

 

IDTS, Aboriginal Male, Alcohol, Moderate
 Subscale  t value  N value  p value
 Unpleasant emotions  3.17  220  p < 0.01
 Physical discomfort  2.80  220  p < 0.01
 Pleasant emotions  3.88  220  p < 0.0001
 Testing personal control  2.35  220  p < 0.05
 Urges/temptations  4.02  220  p < 0.0001
 Conflict with others  3.17  220  p < 0.01
 Social pressure to use  4.72  220  p < 0.0001
 Pleasant times with others  4.72  220  p < 0.0001

 

IDTS, Aboriginal Male, Drugs, Moderate
 Subscale  t value  N value  p value
 Unpleasant emotions  2.47  228  p < 0.05
 Physical discomfort  3.26  228  p < 0.01
 Pleasant emotions  1.75  228  p < 0.10
 Testing personal control  1.46  226  ns
 Urges/temptations  2.35  227  p < 0.05
 Conflict with others  2.47  228  p < 0.05
 Social pressure to use  2.73  227  p < 0.01
 Pleasant times with others  2.73  227  p < 0.01

 

IDTS, Aboriginal Male, Alcohol, High
 Subscale  t value  N value  p value
 Unpleasant emotions  0.57  67  ns
 Physical discomfort  –0.33  67  ns
 Pleasant emotions  –0.14  67  ns
 Testing personal control  –0.57  67  ns
 Urges/temptations  0.20  67  ns
 Conflict with others  0.57  67  ns
 Social pressure to use  1.27  67  ns
 Pleasant times with others  1.27  67  ns

 

IDTS, Aboriginal Male, Drugs, High
 Subscale  t value  N value  p value
 Unpleasant emotions  2.13  83  p < 0.05
 Physical discomfort  1.37  83  ns
 Pleasant emotions  1.64  83  ns
 Testing personal control  –0.19  82  ns
 Urges/temptations  2.48  82  p < 0.05
 Conflict with others  2.13  83  p < 0.05
 Social pressure to use  2.01  82  p < 0.05
 Pleasant times with others  2.01  82  p < 0.05

 

Drug Taking Confidence Questionnaire (DTCQ; Annis, Turner & Sklar, 1997)

The DTCQ measures anticipatory coping self–efficacy as it related to drug and drugs taking risk situations. It specifically assesses a participants belief in his or her ability to cope with potential relapse situations. Items load onto two major types of situations: (a) personal states; and (b) situations involving other people. Under the personal states category there are 5 subscales: (a) unpleasant emotions; (b) physical discomfort; (c) pleasant emotions; (d) testing personal control; and (e) urges/temptations. Under the situations involving other people category, there are three subscales: (a) conflict with others; (b) social pressure to use; and (c) pleasant times with others. It consists of two questionnaires, one for drugs and one for up to three types of drugs.

Results for Non–Aboriginal Male Offenders

DTCQ, Non–Aboriginal Male, Low
 Subscale  t value  N value  p value
 Unpleasant emotions  –1.95  198  p < 0.10
 Physical discomfort  –0.65  198  ns
 Pleasant emotions  –0.97  198  ns
 Testing personal control  –1.24  198  ns
 Urges/temptations  –1.39  198  ns
 Conflict with others  –1.95  198  p < 0.10
 Social pressure to use  –1.85  198  p < 0.10
 Pleasant times with others  –1.85  198  p < 0.10

 

DTCQ, Non–Aboriginal Male, Moderate
 Subscale  t value  N value  p value
 Unpleasant emotions  –20.51  1,218  p < 0.0001
 Physical discomfort  –17.60  1,218  p < 0.0001
 Pleasant emotions  –15.23  1,217  p < 0.0001
 Testing personal control  –16.84  1,217  p < 0.0001
 Urges/temptations  –20.25  1,218  p < 0.0001
 Conflict with others  –20.51  1,218  p < 0.0001
 Social pressure to use  –20.61  1,216  p < 0.0001
 Pleasant times with others  –20.61  1,216  p < 0.0001

 

DTCQ, Non–Aboriginal Male, High
 Subscale  t value  N value  p value
 Unpleasant emotions  –17.52  334  p < 0.0001
 Physical discomfort  –14.44  334  p < 0.0001
 Pleasant emotions  –12.03  334  p < 0.0001
 Testing personal control  –15.44  334  p < 0.0001
 Urges/temptations  –17.70  334  p < 0.0001
 Conflict with others  –17.52  334  p < 0.0001
 Social pressure to use  –18.42  334  p < 0.0001
 Pleasant times with others  –18.42  334  p < 0.0001

Results for Aboriginal Male Offenders

DTCQ, Aboriginal Male, Low
 Subscale  t value  N value  p value
 Unpleasant emotions  –1.93  22  p < 0.10
 Physical discomfort  –1.19  22  ns
 Pleasant emotions  –2.10  22  p < 0.05
 Testing personal control  –2.21  22  p < 0.05
 Urges/temptations  –2.12  22  p < 0.05
 Conflict with others  –1.93  22  p < 0.10
 Social pressure to use  –1.98  22  p < 0.10
 Pleasant times with others  –1.98  22  p < 0.10

 

DTCQ, Aboriginal Male, Moderate
 Subscale  t value  N value  p value
 Unpleasant emotions  –9.31  214  p < 0.0001
 Physical discomfort  –8.49  214  p < 0.0001
 Pleasant emotions  –6.61  214  p < 0.0001
 Testing personal control  –7.84  214  p < 0.0001
 Urges/temptations  –9.49  214  p < 0.0001
 Conflict with others  –9.31  214  p < 0.0001
 Social pressure to use  –9.04  214  p < 0.0001
 Pleasant times with others  –9.04  214  p < 0.0001

 

DTCQ, Aboriginal Male, High
 Subscale  t value  N value  p value
 Unpleasant emotions  –6.47  83  p < 0.0001
 Physical discomfort  –6.41  83  p < 0.0001
 Pleasant emotions  –5.95  83  p < 0.0001
 Testing personal control  –5.52  82  p < 0.0001
 Urges/temptations  –6.62  83  p < 0.0001
 Conflict with others  –6.47  83  p < 0.0001
 Social pressure to use  –7.04  83  p < 0.0001
 Pleasant times with others  –7.04  83  p < 0.0001

 

Effectiveness of Coping Behaviours Inventory (ECBI; Litman, Stapleton, Oppenheim & Peleg, 1984)

The ECBI rates the effectiveness of relapse avoidance methods that abusers employ to avoid high risk situations. The scale does not have subscales but rather one total score.

Results for Non–Aboriginal Male Offenders

 Total Score  t value  N value  p value
 Low  –2.14  172  p < 0.05
 Moderate  –17.29  738  p < 0.0001
 High  –9.81  187  p < 0.0001

Results for Aboriginal Male Offenders

 Total Score  t value  N value  p value
 Low  0.38  28  ns
 Moderate  –9.72  267  p < 0.0001
 High  –5.46  89  p < 0.0001

 

Drug Avoidance Self Efficacy Scale (DASES; Martin, Wilkinson & Poulos, 1995)

The DASES asks respondents to imagine themselves in situations and rate their level of self–efficacy to resist drugs in that situation. The scale does not have subscale but rather one total score.

Results for Non–Aboriginal Male Offenders

 Total Score  t value  N value  p value
 Low  –5.11  282  p < 0.0001
 Moderate  –35.84  1,552  p < 0.0001
 High  –21.01  394  p < 0.0001

Results for Aboriginal Male Offenders

 Total Score  t value  N value  p value
 Low  –2.25  40  p < 0.05
 Moderate  –17.31  354  p < 0.0001
 High  –10.04  108  p < 0.0001

 

Craving Beliefs Questionnaire (CBQ; Beck, Wright, Newman & Liese, 1993)

The CBQ is based on a model that argues that individuals who tend to abuse substances hold three basic types of beliefs: (a) anticipatory (expectation of reinforcement); (b) relief orientation (withdrawal can only be controlled through continued use); and (c) permissive beliefs (rationalizations or excuses). The instrument generally measures commonly held misperceptions about cravings.

Results for Non–Aboriginal Male Offenders

 Total Score  t value  N value  p value
 Low  3.07  196  p < 0.01
 Moderate  23.55  1,349  p < 0.0001
 High  16.41  360  p < 0.0001

Results for Aboriginal Male Offenders

 Total Score  t value  N value  p value
 Low  0.77  26  ns
 Moderate  8.83  268  p < 0.0001
 High  5.94  89  p < 0.0001

 

Beliefs About Substance Abuse (BASA; Beck, Wright, Newman & Liese, 1993)

The BASA is based on a model that argues that individuals who tend to abuse substances hold three basic types of beliefs: (a) anticipatory (expectation of reinforcement); (b) relief orientation (withdrawal can only be controlled through continued use); and (c) permissive beliefs (rationalizations or excuses). The instrument generally measures commonly held misperceptions about substance use.

Results for Non–Aboriginal Male Offenders

 Total Score  t value  N value  p value
 Low  –0.52  199  ns
 Moderate  19.11  1,350  p < 0.0001
 High  11.42  362  p < 0.0001

Results for Aboriginal Male Offenders

 Total Score  t value  N value  p value
 Low  1.55  27  ns
 Moderate  7.51  271  p < 0.0001
 High  6.71  91  p < 0.0001

 

Drinking (and Drugs) Related Locus of Control (DRIE; Donovan & OLeary, 1978)

The DRIE measures an individuals locus of control orientation which is the degree to which personal responsibility is accepted for meaningful events in their lives

Results for Non–Aboriginal Male Offenders

 Total Score  t value  N value  p value
 Low  2.32  264  p < 0.05
 Moderate  21.38  1,414  p < 0.0001
 High  13.31  352  p < 0.0001

Results for Aboriginal Male Offenders

 Total Score  t value  N value  p value
 Low  0.18  37  ns
 Moderate  10.79  338  p < 0.0001
 High  8.62  107  p < 0.0001

 

Paulus Deception Scale (PDS, Paulhus, 1998)

The Paulhus Deception Scale (Paulhus, 1998) measures two constructs: (a) self–deception (the tendency to give self–reports that are honest but positively biased); and (b) impression management (deliberate self–presentation to an audience). The impression management scale was developed under the assumption that some respondents systematically under report their undesirable behaviours, and in extreme cases, respondents are considered to be faking good (i.e. final scores 12 and over). Such styles impact the confidence that can be put on self–report measures, but this does not mean they should be disregarded.

 

Results for non–Aboriginal Male Offenders – Low Intensity

Given that high scores on the impression management domain may be more suspect, it was important to examine the proportion of offender who scored 12 or higher (i.e., may be ‘faking good’). For the impression management scale, pre– and post–program values indicated that only a small percentage of offenders may have been faking good at pre–program (15.41%; N = 55) or post–program (20.60%; N = 55).

 

Results for Aboriginal Male Offenders – Low Intensity

Similar rates emerged for Aboriginal offenders. For the impression management scale, pre– and post–program values indicate only a small percentage of offenders may have been faking good at pre–program (9.09%; N = 5) or post–program (10.26%; N = 4).

 

Results for non–Aboriginal Male Offenders – Moderate Intensity

Given that high scores on the impression management domain may be more suspect, it was important to examine the proportion of offender who scored 12 or higher (i.e., may be faking good). For the impression management scale, pre– and post–program values indicated that only a small percentage of offenders may have been ‘faking good’ at pre–program (7.62%; N = 131) or post–program (13.86%; N = 206).

 

Results for Aboriginal Male Offenders – Moderate Intensity

Similar rates emerged for Aboriginal offenders. For the impression management scale, pre– and post–program values indicate only a small percentage of offenders may have been faking good at pre–program (6.99%; N = 26) or post–program (16.10%; N = 52).

 

Results for non–Aboriginal Male Offenders – High Intensity

Given that high scores on the impression management domain may be more suspect, it was important to examine the proportion of offender who scored 12 or higher (i.e., may be faking good). For the impression management scale, pre– and post–program values indicated that only a small percentage of offenders may have been ‘faking good’ at pre–program (4.38%; N = 21) or post–program (14.78%; N = 56).

 

Results for Aboriginal Male Offenders – High Intensity

Similar rates emerged for Aboriginal offenders. For the impression management scale, pre– and post–program values indicate only a small percentage of offenders may have been ‘faking good’ at pre–program (5.22%; N = 7) or post–program (13.89%; N = 15).

 

Sex Offender Programs

National Low Intensity Sex Offender Pre–/Post–Program Assessment Battery Results

The LI–SOP has a self–report assessment battery to gauge progress toward treatment targets. The assessment battery contains six instruments: (a) Sex Offender Acceptance of Responsibility Scales; (b) Bumby Cognition (Molest and Rape) Scales; (c) Miller Social Intimacy Scale; (d) Empathy Measures; (e) Buss–Perry Aggression Questionnaire; and (f) Balanced Inventory of Desirable Responding. There are small sample sizes (subscale n values range from 6 to 36) with this data and the generalizability of the results may be in question. Given the small sample sizes there was not enough data to disaggregate results into non–Aboriginal and Aboriginal groups; therefore, both these groups are reported together.

Sex Offender Acceptance of Responsibility Scales (SOARS; Peacock, 2000)

The SOARS assesses six dimensions of responsibility for sex offending behaviour: (a) acceptance of sex offence; (b) acceptance of offence planning; (c) acceptance of victim harm; (d) motivation to change; (e) justifications for sexual offending; and (f) acceptance of sexual interests.

Results for SOARS

Results for male offenders on the SOARS revealed that four of the six subscales were significant (p < .05 or better). Specifically, acceptance of sex offence (p < .001), acceptance of offence planning (p < .0001), acceptance of victim harm (p < .01), and acceptance of sexual interests (p < .05) were significant indicating that participants of the low intensity program did accept more responsibility for the offending behaviour.

 

 Subscale  t value  N value  p value
 Acceptance of sex offence  -4.25  22  p < 0.001
 Acceptance of offence planning  -4.93  22  p < 0.0001
 Acceptance of victim harm  -2.94  22  p < 0.01
 Motivation to change  -1.60  22  ns
 Justifications for sexual  offending  -0.68  22  ns
 Acceptance of sexual interests  -2.23  22  p < 0.05

 

Bumby Cognition (Molest and Rape) Scales (BCS; Bumby, 1996)

The BCS is an assessment of cognitive distortions on two subscales of sexual offending, molesting and rape. Higher scores represent greater endorsement of cognitive distortions.

Results for BCS

Results of the BCS indicate significant differences (p < .05 or better) between pre– and post–program indicating that for male offenders, participation in the program appears to have reduced their cognitive distortions around rape and molestation.

 

 Subscale  t value  N value  p value
 Molest  3.43  11  p < 0.01
 Rape  2.22  18  p < 0.05

 

Miller Social Intimacy Scale (MSIS; Miller & Lefcourt, 1982)

The MSIS measures social intimacy and loneliness where higher scores represent higher levels of loneliness.

Results for MSIS

There were no significant differences reported between pre– and post-program scores in the MSIS for the LI–SOP.

 

 Subscale  t value  N value  p value
 Total Score  -0.02  45  ns

 

Empathy Measure (EM; Fernandez, Marshall, Lightbody & OSullivan, 1999)

The EM has three scales, each containing six subscales. The three scales are victim type: (a) child victim; (b) adult woman victim; and (c) adult male victim. Note there was insufficient data to provide analysis on adult male victim scales. Within the victim scales, there are six subscales: (a) recognition of distress toward victim in general; (b) recognition of distress toward victim of sexual assault; (c) recognition of distress toward ones own victim; (d) feelings toward child/women in general; (e) feelings toward child/women as victim of sexual assault; and (f) feelings toward ones own victim.

Results for EM

Results from some of the EM subscales were significant (p < .05 or better) for both victim types. This suggests that participation in the low intensity program may have increased some offender attitudes and beliefs around empathy toward the victim group and their own victims.

 Subscale on Children Victims  t value  N value  p value
 Recognition of distress toward  victim in general  -1.93  37  ns
 Recognition of distress toward  victim of sexual assault  -4.08  37  p < 0.001
 Recognition of distress toward  one’s own victim  -3.54  37  p < 0.001
 Feelings toward child in general  -0.25  37  ns
 Feelings toward child victim of  sexual assault  -0.71  37  ns
 Feelings toward one’s own victim  -1.34  37  ns
 Subscale on Adult Women Victims  t value  N value  p value
 Recognition of distress toward  victim in general  -3.07  7  p < 0.05
 Recognition of distress toward  victim of sexual assault  -3.20  7  p < 0.05
 Recognition of distress toward  one’s own victim  -5.60  7  p < 0.001
 Feelings toward woman as  victim of sexual assault  -2.48  7  p < 0.05
 Feelings toward woman in  general  -0.02  7  ns
 Feelings toward oneâs own  victim  -1.51  7  ns

 

Buss–Perry Aggression Questionnaire (BPAQ; Buss & Perry, 1992)

The BPAQ is a self-report instrument that measures four dimensions of aggression: (a) physical aggression; (b) verbal aggression; (c) anger; and (d) hostility. Scores are expected to decrease with treatment.

Results for BPAQ

There were no significant differences found through the BPAQ. This may reflect that offender attitudes and beliefs around aggression did not change as a result of program participation. However, it should be noted that with small sample sizes, it is difficult to draw any conclusions.

 Subscale  t value  N value  p value
 Physical aggression  -0.08  28  ns
 Verbal aggression  0.61  28  ns
 Anger  -0.16  28  ns
 Hostility  -0.20  28  ns

 

Balanced Inventory of Desirable Responding (BIDR; Paulhus, 1984)

The BIDR measures two constructs: (a) self-deception (the tendency to give self-reports that are honest but positively biased); and (b) impression management (deliberate self-presentation to an audience). The impression management scale was developed under the assumption that some respondents systematically under report their undesirable behaviours, and in extreme cases, respondents are considered to be “faking good” (i.e. final scores 12 and over). Such styles impact the confidence that can be put on self-report measures, but this does not mean they should be disregarded.

Results for BIDR

Given that high scores on the impression management domain may be more suspect, it was important to examine the proportion of offender who scored 12 or higher (i.e. may be “faking good”. For the impression management scale, pre- and post-program values indicate only a small percentage of offenders may have been ‘faking good’ (3.8% (n = 2) at pre-program and 3.8% (n = 2) at post-program).

National Sex Offender Program – Low Intensity Pre–/Post–Program Assessment Battery Results

The LI–SOP has a self–report assessment battery that gauges immediate impacts of program participation through changes in attitudes and beliefs associated with treatment targets such as victim empathy, anger and hostility, accepting responsibility, social intimacy, and cognitive distortions. The assessment battery contains six self–report measures, five of which are used pre–/post–program to gauge treatment performance and one used to assess the authenticity of responses. Within the five scales used pre– and post–program, there were 25 subscales. For the program participants, 12 of the 25 (48%) subscales were statistically significant and in the expected direction at the p < .05 level or better. This indicates that the program participants were reporting some changes in attitudes and beliefs that are consistent with some program targets.

National Moderate Intensity Sex Offender Pre-/Post-Program Assessment Battery Results

The MI–SOP has a self–report assessment battery to gauge progress toward treatment targets. The assessment battery contains six instruments: (a) Sex Offender Acceptance of Responsibility Scales; (b) Bumby Cognition (Molest and Rape) Scales; (c) Miller Social Intimacy Scale; (d) Empathy Measures; (e) Buss-Perry Aggression Questionnaire; and (f) Balanced Inventory of Desirable Responding. There are small sample sizes (subscale n values range from 6 to 36) with this data and the generalizability of the results may be in question. Given the small sample sizes there was not enough data to disaggregate results into non–Aboriginal and Aboriginal groups. Therefore, both these groups are reported together

Sex Offender Acceptance of Responsibility Scales (SOARS; Peacock, 2000)

The SOARS assesses six dimensions of responsibility for sex offending behaviour: (a) acceptance of sex offence; (b) acceptance of offence planning; (c) acceptance of victim harm; (d) motivation to change; (e) justifications for sexual offending; and (f) acceptance of sexual interests.

Results for SOARS

Results for male offenders on the SOARS revealed that three of the six subscales were significant (p < .05 or greater). Specifically, acceptance of sex offence (p < .01), acceptance of offence planning (p < .0001), and acceptance of victim harm (p < .05) were significant indicating that participants of the moderate intensity program did accept more responsibility for the offending behaviour.

 Subscale  t value  N value  p value
 Acceptance of sex offence  -2.79  20  p < 0.01
 Acceptance of offence planning  -3.96  20  p < 0.0001
 Acceptance of victim harm  -2.64  19  p < 0.05
 Motivation to change  -1.97  20  ns
 Justifications for sexual  offending  0.32  20  ns
 Acceptance of sexual interests  0.21  20  ns

 

Bumby Cognition (Molest and Rape) Scales (BCS; Bumby; 1996)

The BCS is an assessment of cognitive distortions on two subscales of sexual offending, molesting and rape. Higher scores represent greater endorsement of cognitive distortions.

Results for BCS

Results of the BCS indicate significant differences (p < .001 or better) between pre– and post–program, indicating that for male offenders, participation in the moderate intensity program appears to have reduced their cognitive distortions around rape and molestation.

 Subscale  t value  N value  p value
 Molest  4.38  34  p < 0.0001
 Rape  4.28  30  p < 0.001

 

Miller Social Intimacy Scale (MSIS; Miller & Lefcourt, 1982)

The MSIS measures social intimacy and loneliness where higher scores represent higher levels of loneliness.

Results for MSIS

There were no significant differences reported between pre- and post-program scores in the MSIS for the MI-SOP.

 Subscale  t value  N value  p value
 Total score  -0.93  91  ns

 

Empathy Measure (EM; Fernandez, Marshall, Lightbody & O’Sullivan, 1999)

The EM has three scales each containing six subscales. The three scales are victim type: (a) child victim; (b) adult woman victim; and (c) adult male victim. Note there was insufficient data to provide analysis on adult male victim scales. Within the victim scales, there are six subscales: (a) recognition of distress toward victim in general; (b) recognition of distress toward victim of sexual assault; (c) recognition of distress toward one’s own victim; (d) feelings toward child/women in general; (e) feelings toward child/women as victim of sexual assault; and (f) feelings toward one’s own victim.

Results for EM

Results from some of the EM subscales were significant (p < .05 or better) for both victim types (women and child victims). This suggests that participation in the moderate intensity program may have increased some offender attitudes and beliefs around empathy toward the victim group and their own victims.

 Subscale on Children Vitims  t value  N value  p value
 Recognition of distress toward  victim in general  -2.09  57  p < 0.05
 Recognition of distress toward  victim of sexual assault  -2.66  55  p < 0.01
 Recognition of distress toward  one’s own victim  -4.28  53  p < 0.0001
 Feelings toward child in general  -1.29  56  ns
 Feelings toward child victim of  sexual assault  -2.13  55  p < 0.05
 Feelings toward one’s own victim  -2.26  54  p < 0.05
 Subscale on Adult Women  Victims  t value  N value  p value>
 Recognition of distress toward  victim in general  -3.32  40  p < 0.01
 Recognition of distress toward  victim of sexual assault  -2.15  40  p < 0.05
 Recognition of distress toward  one’s own victim  -2.96  39  p < 0.01
 Feelings toward woman as  victim of sexual assault  -0.11  39  ns
 Feelings toward woman in  general  -0.86  40  ns
 Feelings toward one’s own  victim  -3.20  39  p < 0.01

 

Buss–Perry Aggression Questionnaire (BPAQ; Buss & Perry, 1992

The BPAQ is a self-report instrument that measures four dimensions of aggression: (a) physical aggression; (b) verbal aggression; (c) anger; and (d) hostility. Scores are expected to decrease with treatment.

Results for BPAQ

Results from the BPAQ reveal significant differences pre- and post-program. Specifically, there were significant differences on the verbal aggression (p < .01), anger (p < .001), and hostility (p < .05) subscales. These results suggest that the MI-SOP program participants may have changed some attitudes and beliefs around anger and hostility.

 Subscale  t value  N value  p value
 Physical aggression  -1.29  76  ns
 Verbal aggression  2.86  76  p < 0.01
 Anger  3.52  76  np < 0.001
 Hostility  -2.10  76  p < 0.05

 

Balanced Inventory of Desirable Responding (BIDR; Paulhus, 1984)

The BIDR measures two constructs: (a) self-deception (the tendency to give self-reports that are honest but positively biased); and (b) impression management (deliberate self-presentation to an audience). The impression management scale was developed under the assumption that some respondents systematically under report their undesirable behaviours, and in extreme cases, respondents are considered to be “aking good” (i.e. final scores 12 and over). Such styles impact the confidence that can be placed on self-report measures, but this does not mean they should be disregarded.

 

Results for BIDR

Given that high scores on the impression management domain may be more suspect, it was important to examine the proportion of offender who scored 12 or higher (i.e. may be ‘faking good’). For the impression management scale, pre- and post-program values indicate only a small percentage of offenders may have been ‘faking good’ (3.1% (n = 4) at pre-program and 0.8% (n = 1) at post-program).

National Sex Offender Program – Low Intensity Pre–/Post-Program Assessment Battery Results

The LI–SOP has a self-report assessment battery that gauges immediate impacts of program participation through changes in attitudes and beliefs associated with treatment targets such as victim empathy, anger and hostility, accepting responsibility, social intimacy, and cognitive distortions. The assessment battery contains six self-report measures, five of which are used pre- post-program to gauge treatment performance and one used to assess the authenticity of responses. Within the five scales used pre- and post-program, there were 25 subscales. For the program participants, 17 of the 25 (68%) subscales were statistically significant and in the expected direction at the p < .05 level or greater. This indicates that the program participants were reporting some changes in attitudes and beliefs that are consistent with some program targets.

Living Skills Programs

Anger and Emotions Management Pre–/Post-Program Assessment Battery

The assessment battery for Anger and Emotions Management (A & EM) contained 5 assessment instruments that were analysed: (a) Revised Anger and Emotions Management Questionnaire; (b) Reactions to Provocation; (c) Reduced Emotional Intensity Scale; (d) Barratt Impulsivity Scale; and (e) Balanced Inventory of Desirable Responding. Older versions of the assessment battery contained additional scales, but only scales contained in the current battery were examined.

 

Revised Anger and Emotions Management Questionnaire (RAEMQ; CSC, Reintegration Programs Branch 1993)

The RAEMQ is a 38–item self–report measure completed pre– and post–program assessing a wide variety of indicators of anger management skills. The instrument contains three subscales, each assessing specific aspects of anger: (a) insight into anger; (b) knowledge and anger management skills; and (c) anger management self-competence. Scores are expected to increase with intervention.

Reliability and Validity

Change scores on the three subscales were associated with significant reductions in violent recidivism indicating predictive validity (Dowden et al., 1999).

Results for non–Aboriginal Male Offenders

As expected, all subscale scores increased significantly as a result of participation in the program (p < .0001). Results seem to indicate that offenders are self–reporting better anger management self competencies and insights into their anger.

 Subscale  t value  N value  p value
 Insight  -22.68  337  p < 0.0001
 Knowledge  -17.80  338  p < 0.0001
 Anger management  -17.97  338  p < 0.0001

Results for Aboriginal Male Offenders

The same results emerged for Aboriginal offender than what emerged for non–Aboriginal offenders. These results also seem to indicate that Aboriginal offenders are self–reporting better anger management self competencies and insights into their anger.

 Subscale  t value  N value  p value
 Insight  -6.93  71  p < 0.0001
 Knowledge  -6.81  71  p < 0.0001
 Anger management  -6.73  71  p < 0.0001

 

Reactions to Provocations (RP; Novaco, 1994)

The RP is a 73–item self-report measure completed pre– and post–program assessing the behavioural response associated with the cognitive and arousal elements of anger. The instrument contains four subscales: (a) cognitive (hostility, suspiciousness and rumination); (b) arousal (intensity and duration of physiological arousal); (c) behavioural (verbally aggressive and confrontational); and (d) intensity of anger (disposition to be disrespectful and to become frustrated and annoyed). Scores are expected to decrease with intervention.

Reliability and Validity

Reliability was indicated through high test-retest correlations (exceeding.80; Mills, Kroner & Forth, 1998). The subscales demonstrated a stronger relationship with their respective domains than with other domains indicating construct validity. There was also a strong relationship found between this scale and other anger and aggression measures, providing evidence for concurrent validity.

Results for non-Aboriginal Male Offenders

As expected, scores were significantly reduced (p < .0001) post-program for each of the four subscales. These results indicate that offenders have fewer deficits in thinking patterns, a reduced intensity and duration of physiological arousal, and a reduced tendency to react impulsively.

 Subscale  t value  N value  p value
 Cognitive  14.00  338  p < 0.0001
 Arousal  13.94  338  p < 0.0001
 Behavioural  15.28  338  p < 0.0001
 Intensity of anger  11.55  337  p < 0.0001

Results for Aboriginal Male Offenders

The same results emerged for Aboriginal offenders. These results indicate that offenders have fewer deficits in thinking patterns, a reduced intensity and duration of physiological arousal, and a reduced tendency to react impulsively.

 Subscale  t value  N value  p value
 Aspects cognitifs  5,72  71  p < 0,0001
 Excitation  4,96  71  p < 0,0001
 Aspects comportementaux  5,83  71  p < 0,0001
 Intensité de la colère  5,76  71  p < 0,0001

 

Reduced Emotional Intensity Scale (REIS; Geuens & DePelsmacker, 2002)

The REIS is a 17–item self-report measure completed pre- and post-program to assess the intensity of emotions and distinguished between negative and positive emotions. The scale has two subscales: (a) positive emotional intensity; and (b) negative emotional intensity. For the positive emotional intensity subscale, scores are expected to increase with intervention. For the negative emotional intensity subscale, scores are expected to decrease.

Reliability and Validity

The instrument has demonstrated test-retest reliability where correlations for all subscales were above .80 (Geuens & DePelsmacker, 2002). Construct validity was also demonstrated by examining the correlations of the scores of both factors with scores on other traits such as neuroticism, extraversion, and affect intensity.

Results for Non-Aboriginal Offenders

Results indicated that offenders had significant changes on both subscales. A significant reduction in positive emotional intensity (p < .05) suggests an increase in positive affect while a significant reduction in negative emotional intensity (p < .0001) suggests a reduction in the presence of negative emotions such as anxiety, tension, and nervousness.

 Subscale  t value  N value  p value
 Positive emotional intensity  2.05  337  p < 0.05
 Negative emotional intensity  4.96  71  p < 0.0001

Results for Aboriginal Offenders

Results indicated that Aboriginal offenders had significant changes on only one of the two subscales. A significant reduction in negative emotional intensity (p < .0001) suggests a reduction in the presence of negative emotions such as anxiety, tension, and nervousness.

 Subscale  t value  N value  p value
 Positive emotional intensity  0.26  71  ns
 Negative emotional intensity  3.87  71  p < 0.001

 

Barratt Impulsivity Scale (BIS; Barrett, 1994)

The BIS is a 30-item self-report measure completed pre- and post-program assessing impulsivity. Research has suggested that impulsivity can be viewed as three subtraits: (a) motor impulsivity (to act without thinking); (b) attentional impulsivity (making quick cognitive decisions); and (c) non–planning (lack of concern for the future). Scores are expected to lower with intervention.

Reliability and Validity

Factor analysis comparisons between those with psychopathology and those without indicated that individuals with impulsive aggressive tendencies tended to score higher than matched controls in the three areas (Monahan & Steadman, 1994).

Results for Non– Aboriginal Offenders

Results found significant reductions on all three subscales (p < .0001). These results indicate that as a result of program participation, offenders were more reluctant to act on the spur of the moment without thinking, and they increased consideration given to decision making and solving problems with greater forethought.

 Subscale  t value  N value  p value
 Motor impulsivity  12.59  335  p < 0.0001
 Attentional impulsivity  11.09  335  p < 0.0001
 Non-planning impulsivity  17.85  402  p < 0.0001

Non–planning impulsivity

The same results were found for Aboriginal offenders. These results indicate that as a result of program participation, offenders were more reluctant to act on the spur of the moment without thinking, and they increased consideration given to decision making and solving problems with greater forethought.

 Subscale  t value  N value  p value
 Motor impulsivity  4.55  71  p < 0.0001
 Attentional impulsivity  4.83  71  p < 0.0001
 Non-planning impulsivity  9.75  92  p < 0.0001

 

Balanced Inventory of Desirable Responding (BIDR; Paulhus, 1984)

The BIDR is 40-item self-report measure completed only at pre-program that measures two constructs, self-deception (the tendency to give self-reports that are honest but positively biased), and impression management (deliberate self-presentation to an audience). The impression management scale was developed under the assumption that some respondents systematically under report their undesirable behaviours, and in extreme cases, respondents are considered to be “faking good” (i.e. final scores 12 and over). Such styles impact the confidence that can be put on self-report measures, but this does not mean they should be disregarded.

Reliability and Validity

The internal reliability or the extent to which the scale consistently measures the same construct is highly satisfactory. Both subscales have been found to have good convergent validity.

Results for non-Aboriginal Male Offenders

Given that high scores on the impression management domain may be more suspect, it was important to examine the proportion of offender who scored 12 or higher (i.e. may be “faking good”). For the impression management scale, pre- and post-program values indicate only a small percentage of offenders may have been ‘faking good’ 8.7% (n = 39) at pre-program.

Results for Aboriginal Male Offenders

Similar rates emerged for Aboriginal offenders. For the impression management scale, pre- and post-program values indicate only a small percentage of offenders may have been “faking good” 7.5% (n = 7) at pre-program.

Reasoning and Rehabilitation Pre-/Post-Program Assessment Battery

The assessment battery for R & R contained three assessment instruments that were analysed: (a) Social Problem Solving Inventory – Revised Short Form; (b) Barratt Impulsivity Scale; and (c) Motivation to Change Index. Older versions of the assessment battery contained additional scales, but only scales contained in the current battery were examined.

 

Social Problem Solving Inventory – Revised Short Form (SPSR; D’Zurilla, Nezu & Maydeu-Olivares, 1999)

The SPSR is a 25-item self-report measure completed pre– and post–program assessing offenders’ attitudes toward solving problems. The social problem–solving process contains five subscales: (a) positive problem orientation (constructive problem solving); (b) negative problem orientation (dysfunctional attitude toward problem solving); (c) rational problem solving (rational and effective problem solving techniques); (d) impulsive/careless style (careless and incomplete problem solving); and (e) avoidant style (procrastination, passivity or inaction).

Reliability and Validity

No reliability or validity information is available on this instrument.

Results for Non–Aboriginal Offenders

Two subscales are expected to increase (positive problem orientation and rational problem solving) and three are expected to go down (negative problem solving, impulsive/ careless problem solving, and avoidant style). Results for non-Aboriginal offenders reveal that all subscales indicate that offenders are self-reporting more effective problem solving styles post-program in comparison to pre-program scores. This suggests that offenders, through participation in the program, changed some attitudes and beliefs regarding their social problem solving techniques.

 Subscale  t value  N value  p value
 Positive problem orientation  -11.78  562  p < 0.0001
 Negative problem orientation  9.32  563  p < 0.0001
 Rational problem solving  -16.53  559  p < 0.0001
 Impulsive/careless problem solving  15.16  556  p < 0.0001
 Avoidant style  12.28  562  p < 0.0001

Results for Aboriginal Offenders

The exact same results emerged for Aboriginal offenders. All subscales indicate that offenders are self-reporting more effective problem solving styles post-program in comparison to pre-program scores. This suggests that Aboriginal offenders, through participation in the program, changed some attitudes and beliefs regarding their social problem solving techniques.

 Subscale  t value  N value  p value
 Positive problem orientation  -5.55  108  p < 0.0001
 Negative problem orientation  3.93  108  p < 0.001
 Rational problem solving  -7.64  108  p < 0.0001
 Impulsive/careless problem solving  5.71  108  p < 0.0001
 Avoidant style  4.92  108  p < 0.0001

 

Barratt Impulsivity Scale (BIS; Barrett, undated)

The BIS is a 30-item self-report measure completed pre- and post-program assessing impulsivity. Research has suggested that impulsivity can be viewed as three subtraits: (a) motor impulsivity (to act without thinking); (b) attentional impulsivity (making quick cognitive decisions); and (c) non-planning (lack of concern for the future). Each of these three subtraits form a subscale. Scores are expected to lower with intervention.

Reliability and Validity

Factor analysis comparisons between those with psychopathology and those without indicated that individuals with impulsive aggressive tendencies tended to score higher than matched controls in the three areas (Monahan & Steadman, 1994).

Results for Non–Aboriginal Offenders

Results indicate that for motor impulsivity and non-planning impulsivity subscales offenders were self-reporting significantly less impulsivity in their responses (p < .0001). However, for the attentional subscale, there was also significance found (< .0001) but the direction is opposite to what was expected indicating non-Aboriginal offenders are self-reporting more attentional impulsivity after completion of the program. This suggests that non-Aboriginal offenders, through participation in the program, changed some attitudes and beliefs regarding impulsivity and may have learned some mechanisms to reduce their impulsivity.

 Subscale  t value  N value  p value
 Motor impulsivity  15.25  538  p < 0.0001
 Attentional impulsivity  -4.06  538  p < 0.001
 Non-planning impulsivity  16.16  542  p < 0.0001

Results for Aboriginal Offenders

The same pattern emerged for Aboriginal offender data. Results indicate that for motor impulsivity and non-planning impulsivity subscales Aboriginal offenders were self-reporting significantly less impulsivity in their responses (p < .0001). However, for the attentional subscale, there was also significance found (p < .0001) but the direction is opposite to what was expected indicating offenders are self-reporting more attentional impulsivity after completion of the program. This suggests that Aboriginal offenders, through participation in the program, changed some attitudes and beliefs regarding impulsivity and may have learned some mechanisms to reduce their impulsivity.

 Subscale  t value  N value  p value
 Motor impulsivity  5.41  106  p < 0.0001
 Attentional impulsivity  -3.92  105  p < 0.001
 Non-planning impulsivity  6.57  104  p < 0.0001

 

Motivation to Change Index (MCI)

The MCI questionnaire is completed by the interviewer at pre- and post-program through a semi-structured interview with the offender. The interviewer rates the offenders’ answers and provides a rating for each question and a total stage score. The MCI is based on the Stages of Change Model where the interviewer is able to assess the extent to which the offender recognises the deficit as a problem and is willing to take action to change it by participating in the program. The interview will make a statement (i.e., I usually don’t stop and think before acting) and ask the offender which of the statement are true about him and how the problem may affect him. The offender is rated on their response by placement into a scale from pre-contemplation to maintenance. The instrument measures socio-cognitive deficit areas and provides an overall stages of change estimate. Based of the social-cognitive deficit ratings, an overall stage estimate is made pre- and post-program. Scores are expected to increase with intervention.

Reliability and Validity

No reliability or validity information is available on this instrument.

Results for Non–Aboriginal Offenders

As expected, results reveal that the overall stage of change estimate has increased from pre– to post-program indicating that non–Aboriginal offenders have become more motivated to change through participation in the program and have moved through the stages of change.

 Subscale  t value  N value  p value
 Overall stage of change  estimate  -19.26  507  p < 0.0001

Results for Aboriginal Offenders

As expected, results reveal that the overall stage of change estimate has increased from pre– to post–program indicating that Aboriginal offenders have become more motivated to change through participation in the program and have moved through the stages of change.

 Subscale  t value  N value  p value
 Overall stage of change  estimate  -4.39  92  p < 0.0001

 

Appendix I: Staff Responses & Recommendations

Table I1. Staff Recommendations to Improve Efficiency of OIA Process
 Recommendations:  Staff Respondents N (%)
 More training/certification for intake staff  24 (14.6%)
 Increase the amount of resources  20 (12.2%)
 Reduce/eliminate the use of self-report data  19 (11.6%)
 Increase time allotted to complete the OIA  12 (7.3%)
 Decrease the time taken to obtain community reports  10 (6.1%)
 More teamwork/cooperation throughout the intake process  10 (6.1%)
 Decrease the number of assessments/instruments  9 (5.5%)
 Greater use of offender background information/histories  7 (4.3%)
 Have more experienced staff administer the OIA  7 (4.3%)
 Increase the accuracy/detail of assessments  7 (4.3%)
 Adjust timing of OIA to avoid overwhelming offenders  7 (4.3%)
 More communication between intake staff and CPOs/POs  7 (4.3%)
 Shorter assessment periods for shorter sentences  6 (3.7%)
 More clarity in program decision/placement rationale  6 (3.7%)
 Decrease ability to override assessment results  6 (3.7%)
 Have CPOs working at intake  6 (3.7%)
 More psychological assessments/information  5 (3.0%)
 Increase accuracy of risk assessments  5 (3.0%)
 More communication between sites (intake, facilities, and community)  5 (3.0%)
 Ensure all assessments completed before placed in facility  5 (3.0%)
 Centralise the intake process  5 (3.0%)
 More one-on-one interviews  4 (2.4%)
 Dedicated resources for Aboriginal offenders at intake  3 (1.8%)
 Do not record insignificant assessment findings in OMS  3 (1.8%)
 Offer programs to offenders during OIA  2 (1.2 %)
 Update the OIA to be more current  2 (1.2%)
 Make the OIA more applicable to women  2 (1.2%)
 Improve offender movement at intake  2 (1.2%)
 More computer automation to decrease multiple data entry  2 (1.2%)
 More emphasis on inmates mother tongue  1 (< 1%)
 Greater access to psych reports  1 (< 1%)
 Target supplementary assessments to those at risk  1 (< 1%)
 Increase accuracy of community reports  1 (< 1%)
 Conduct supplementary assessments earlier in the OIA  1 (< 1%)
 Stop over-programming/address more pressing need only  1 (< 1%)
 Access to juvenile records  1 (< 1%)
 Decrease use of Elders/more standardised assessments for Aboriginal offenders  1 (< 1%)
 Do intake at remand centres  1 (< 1%)
 OIA should be 12 hours/day (less downtime)  1 (< 1%)
Note: Due to multiple responses, percentages may sum to greater than 100%.

 

Table I2. Outstanding Offender Information Required by Staff
 Information Required:  Staff Respondents N (%)
 More detailed offender information (psychiatric reports, histories)  20 (31.7%)
 Police reports/court documents  13 (20.6%)
 Program start dates/progress/completions  6 (9.5%)
 Medical information  5 (7.9%)
 Provincial information (programs, police records)  4 (6.3%)
 Security/institutional behaviour information  4 (6.3%)
 User-friendly access to reports/information  3 (4.8%)
 Access to shared network drive for contractors  1 (1.6%)
 Elder assessments  1 (1.6%)
 Parole officers’ program recommendations  1 (1.6%)
 Contact with offender’s family and friends  1 (1.6%)
 Education transcripts  1 (1.6%)
 Information for international transfers in OMS  1 (1.6%)
 Rationale for overriding program referrals  1 (1.6%)
Note. 12.7% of staff responses (N = 8) did not reflect the topic area in question.

 

Table I3. Staff Suggestions for Improving the Usefulness and/or Efficiency of Supplementary Intake Assessments
 Suggestions  Répondants au sein du personnel n (%)
 Staff Respondents N (%)  40 (29.2%)
 Increase objectivity / validity of assessments  30 (21.9%)
 Improve staff training / ensure assessor is sufficiently trained  24 (17.5%)
 Allow for overrides of assessment results with clinical judgment  23 (16.8%)
 Develop new assessment tools / expand upon existing tools  19 (13.9%)
 Administer in-person instead of using a computer  7 (5.1%)
 Reassess offenders regularly / more than once  6 (4.4%)
 Incorporate specific responsivity factors into assessments  5 (3.6%)
 Make assessment tools easier / less complicated  4 (2.9%)
 Do not incorporate assessment of risk  3 (2.2%)
 Assist offenders with completing self-report assessments  3 (2.2%)
 Increase offender awareness of purpose of assessment  3 (2.2%)
 Ensure timely access to information required to complete assessments  3 (2.2%)
 Accommodate offenders who are not computer literate  2 (1.5%)
 Increase assessment staff / resources  2 (1.5%)
 Assess all offenders with the CLAI/CASA  2 (1.5%)
 Link assessments to data in OMS  2 (1.5%)
 Ensure high-scoring offenders are placed in high intensity programs  1 (<1%)
 Improve cultural sensitivity of assessment tools  1 (<1%)
 Use FVRA/SARA with women offenders  1 (<1%)
Note: Due to multiple responses, percentages may sum to greater than 100%.

 

Table I4. Reasons provided by staff as to why an offender might not participate in a program listed in his or her correctional plan
 Reasons  Staff Respondents N (%)
 Offender’s refusal to participate  153 (65.4%)
 Short sentence length / lack of time to complete program  85 (36.3%)
 Lack of available programs at the institution  84 (35.9%)
 Specific responsivity factors (mental illness, cognitive impairment, low educational attainment, etc.)  52 (22.2%)
 Timing of program delivery at the institution  45 (19.2%)
 Long waiting list to take the program  43 (18.4%)
 Case under appeal / offender denies responsibility  32 (13.7%)
 Offender feels program is unnecessary  31 (13.2%)
 Incorrect or inappropriate referral  30 (12.8%)
 Lack of available programs in the community  28 (12.0%)
 Fear of labelling / stigma associated with program  25 (10.7%)
 Conflict with employment / employment is priority  24 (10.3%)
 Offender placed in segregation / protective custody / lockdown / offender movement issues  18 (7.7%)
 Insufficient staff / program facilitators  17 (7.3%)
 Other program takes priority  16 (6.8%)
 Discomfort in group / incompatibles in group  15 (6.4%)
 Timing of program delivery in the community  14 (6.0%)
 Program not offered in preferred language  12 (5.1%)
 Offender’s security classification does not correspond with intensity level of required program  10 (4.3%)
 Offender refuses to transfer to program delivery site  10 (4.3%)
 Conflict with program facilitator  7 (3.0%)
 Change in referral criteria / no longer meets referral criteria  7 (3.0%)
 Offender disagrees with assessment / referral  6 (2.6%)
 Peer pressure to refuse participation / drop out  6 (2.6%)
 Parole Officer disagrees with program referral  5 (2.1%)
 Not enough offenders who meet referral criteria to form a group  4 (1.7%)
 Program is not culturally sensitive to Aboriginal offenders  3 (1.3%)
 No longer needs program in community  3 (1.3%)
 Offender fears participation in program  2 (<1%)
 Offender in court for extended period of time  2 (<1%)
 Lack of transportation to community program  2 (<1%)
 Did not complete prerequisite program  1 (<1%)
 Reincarcerated while participating in community program  1 (<1%)
 Limit on the number of high intensity programs to which an offender can be referred  1 (<1%)
Note: Due to multiple responses, percentages may sum to greater than 100%.

 

Table I5. Recommendations for Improving Correctional Program Outcomes for Offenders with Mental Health Needs
 Recommendations:  Staff Respondents N (%)
 Create new programs for those with mental health needs  71 (38.2%)
 More 1-on-1 interventions/psych services  30 (16.1%)
 More training/skills for staff  23 (12.4%)
 Adjust currents programs to accommodate needs  22 (11.8%)
 More resources (funding, psychologists)  22 (11.8%)
 More identifying assessments  16 (8.6%)
 More community support/follow-up  16 (8.6%)
 More team-oriented approach  13 (7.0%)
 More communication regarding offender mental health  11 (5.9%)
 More flexible program delivery schedule  10 (5.4%)
 New pen/system for mentally ill offenders  8 (4.3%)
 More mental health interventions  7 (3.8%)
 Smaller group sizes for these offenders  5 (2.7%)
 More research on mental health issues  4 (2.2%)
 More assistants in groups to help these offenders  4 (2.2%)
 Increase the capacity of mental health units  3 (1.6%)
 Less reliance on medicating offenders  3 (1.6%)
 More access to existing mental health programs  1 (<1%)
 More institutional support groups  1 (<1%)
 More booster/maintenance programs  1 (<1%)
Note: Several respondents (4.3%, N = 8) provided answers unrelated to the question; these responses are not listed.

 

Table I6: Recommendations for Improving Correctional Program Outcomes for Offenders with Special Needs
 Recommendations:  Staff Respondents N (%)
 Implement new programs specific to group/need  57 (29.4%)
 Adapt current programs to meet needs  49 (25.3%)
 More staff training  40 (20.6%)
 More 1-on-1 programming  21 (10.8%)
 More tutors/teacher assistants/specialists  21 (10.8%)
 More flexible program schedules  20 (10.3%)
 More identifying/assessment tools  19 (9.8%)
 More resources  19 (9.8%)
 Smaller group sizes  9 (4.6%)
 More education/ESL for these offenders  9 (4.6%)
 More community support/programs  5 (2.6%)
 More communication with staff/health services  4 (2.1%)
 More individualized treatment plans  4 (2.1%)
 More follow-up/supports  3 (1.5%)
 More mentoring  3 (1.5%)
 More appropriate/therapeutic environment  3 (1.5%)
 More team-orientated approach  2 (1.0%)
 Move to mental health facility  2 (1.0%)
 More behavioural management efforts  1 (<1%)
 Professionalize the CPO position  1 (<1%)
Note: Several respondents (3.6%, N = 7) provided answers unrelated to the question; these responses are not listed.

 

Table I7. Potential impact on correctional programs of legislation concerning the elimination of statutory release
 Impacts:  Staff Respondents N (%)
 Offenders will be more motivated to participate in programs  104 (48.8%)
 Offenders will be less motivated to participate in programs  56 (26.3%)
 Offenders will have a longer amount of time in the institution in which to participate in programs  53 (24.9%)
 Greater demand/need for programs in institutions  37 (17.4%)
 Increased need for resources within institutions  18 (8.5%)
 Decreased availability of programs in community  16 (7.5%)
 Increase in incarcerated population / overcrowding  15 (7.0%)
 Participation will be delayed further into offender’s sentence  13 (6.1%)
 Change in how offenders are assigned to programs  7 (3.3%)
 Longer wait lists  6 (2.8%)
 Increased need for diversified programs  6 (2.8%)
 May create legal issues for institutions  3 (1.4%)
 Increased cost to Canada  3 (1.4%)
 Not enough time to complete programs  3 (1.4%)
 Programs will better target offender  2 (<1%)
 Increase offender accountability  2 (<1%)
 Will impact start date of programs  2 (<1%)
 Increase success of programs  2 (<1%)
 Longer sentences will deter offenders and lead to a lesser demand for programs  2 (<1%)
 Don’t know/Not sure  2 (<1%)
Note: Twelve respondents (5.6%) provided responses that were unrelated to the question.

 

Table I8. Recommendations for Improving Correctional Program Outcomes for Aboriginal Offenders
 Recommendations:  Staff Respondents N (%)
 More/greater range of Aboriginal specific programming  60 (37.5%)
 More Aboriginal content in existing programming  27 (16.9%)
 More resources (Elders/ALOs/therapists)  23 (14.4%)
 More staff training on Aboriginal culture  19 (11.9%)
 More Aboriginal staff (CPOs/POs)  18 (11.3%)
 More contact with facilitators/Elders  15 (9.4%)
 More community liaising/connections  15 (9.4%)
 Greater use of healing plans/Elder assessments  13 (8.1%)
 More structure/accreditation in Aboriginal programs  8 (5.0%)
 More Elder training on correctional approach/reports  8 (5.0%)
 More use of Aboriginal learning/teaching styles  7 (4.4%)
 More knowledge of Aboriginal resources  7 (4.4%)
 More programs in the community  6 (3.8%)
 More programs dealing with Aboriginal abuse/FAS  5 (3.1%)
 More encouragement toward spirituality  4 (2.5%)
 Separate location for Aboriginals/Aboriginal range  3 (1.9%)
 More Aboriginal programming in minimum security  3 (1.9%)
 More teamwork/communication between CPOs and Elders  2 (1.3%)
 More jobs involving cultural activities  2 (1.3%)
 Decrease minimum # of Aboriginal offenders needed to run programs  2 (1.3%)
 Less reliance on self-report  1 (<1%)
 Ensure programs actually addressing need  1 (<1%)
Note: Several respondents (8.1%, N = 13) provided answers unrelated to the question; these responses are not listed.

 

Table I9. Recommendations for Improving Correctional Program Outcomes for Women Offenders
 Recommendations:  Staff Respondents N (%)
 More/greater variety of women specific programming  28 (46.7%)
 More community contact/ETAs  6 (10.0%)
 More community programs  6 (10.0%)
 More consideration of motherhood in programs  6 (10.0%)
 Treat them the same as male offenders  4 (6.7%)
 Improve women’s risk assessments  4 (6.7%)
 More staff/resources  3 (5.0%)
 More vocational programs/training  3 (5.0%)
 More Aboriginal women offender programs  2 (3.3%)
 More training for staff to understand needs of women  2 (3.3%)
 More holistic approach  2 (3.3%)
 More communication between CMT and Psychologist  1 (1.7%)
 More research-based programs  1 (1.7%)
 Less babying/coddling  1 (1.7%)
 More flexibility in scheduling (due to family responsibilities)  1 (1.7%)
 More one-on-one interventions  1 (1.7%)
 More focus on those with short sentences  1 (1.7%)
Note: Several respondents (10.0%, N = 6) provided answers unrelated to the question; these responses are not listed.

 

Table I10. Correctional Programs staff respondents were trained to deliver
 Correctional Program:  Staff Respondents N (%)
 National Substance Abuse Program – High Intensity  58 (43.9%)
 National Substance Abuse Program – Moderate Intensity  55 (41.7%)
 National Substance Abuse Program – Low Intensity  55 (41.7%)
 National Substance Abuse Program Maintenance  46 (34.8%)
 Offender Substance Abuse Program / Choices  18 (13.6%)
 Women Offender Substance Abuse Program  14 (10.6%)
 Aboriginal Offender Substance Abuse Program  10 (7.6%)
 Substance Abuse Program Booster  3 (2.3%)
 Methadone  2 (1.5%)
 High Intensity Family Violence Prevention Program  24 (18.2%)
 Moderate Intensity Family Violence Prevention Program  24 (18.2%)
 Family Violence Maintenance  17 (12.9%)
 Aboriginal High Intensity Family Violence Prevention Program  5 (3.8%)
 Living without Violence  3 (2.3%)
 National Sex Offender Program – Moderate Intensity  23 (17.4%)
 National Sex Offender Program – Low Intensity  23 (17.4%)
 National Sex Offender Program Maintenance  22 (16.7%)
 National Sex Offender Program – High Intensity  14 (10.6%)
 Discontinued Sex Offender Program  3 (2.3%)
 Local Sex Offender Program  2 (1.5%)
 Aboriginal Sex Offender Program  1 (<1%)
 Violence Prevention Program – High Intensity  11 (8.3%)
 Violence Prevention Program – Moderate Intensity  11 (8.3%)
 Violence Prevention Program Maintenance  8 (6.1%)
 In Search Of Your Warrior  6 (4.5%)
 Women Violence Prevention Program  3 (2.3%)
 Spirit of A Warrior  2 (1.5%)
 Discontinued Violence Prevention Program (Unspecified)  1 (<1%)
 Community Maintenance Program  17 (12.9%)
 Community Reintegration  7 (5.3%)
 Unspecified Community Maintenance Program  1 (<1%)
 Spirit of A Warrior Maintenance  1 (<1%)
 Reasoning & Rehabilitation / Cognitive Skills  56 (42.4%)
 Anger & Emotions Management  44 (33.3%)
 Cognitive Skills Maintenance  29 (22.0%)
 Anger & Emotions Management Maintenance  24 (18.2%)
 Parenting Skills  6 (4.5%)
 Aboriginal Basic Healing Program  5 (3.8%)
 Leisure Skills  2 (1.5%)
 Life skills  2 (1.5%)
 Cognitive Skills Booster  1 (<1%)
 Attitudes, Alternatives & Associates  23 (17.4%)
 Counter-Point  10 (7.6%)
 Circles of Change  4 (3.0%)
 National Segregation Program  2 (1.5%)
 Gambling  1 (<1%)
 Conflict Resolution  1 (<1%)
 Aboriginal Parenting  1 (<1%)
 Self-Monitoring  1 (<1%)
 Other Unspecified Program  1 (<1%)

 

Table I11. Reasons listed by staff for not delivering a program in which they were trained to deliver
 Reasons  Staff Respondents N (%)
 Program not offered  7 (21.2%)
 Changed jobs/institutions  3 (9.1%)
 Personal choice  3 (9.1%)
 Program was assigned to another CPO  3 (9.1%)
 Training used for other purposes (e.g., professional development)  3 (9.1%)
 Newly trained / no opportunity to deliver program  3 (9.1%)
 Program was cancelled  2 (6.1%)
 Trained as back-up  2 (6.1%)
 Was given other tasks/priorities within institution  2 (6.1%)
 Low priority for institution  2 (6.1%)
 Program not yet translated  1 (3.0%)
 Was not required to deliver entire program  1 (3.0%)
 Program has not yet started  1 (3.0%)
Note: Two respondents (6.1%) provided suggestions unrelated to the question.

 

Table I12. Potential impacts of insufficient numbers of trained CPOs, as provided by staff
 Responses  Staff Respondents N (%)
 Offenders are released without program/intervention  24 (26.7%)
 Long waiting lists  15 (16.7%)
 Delay in release/parole  15 (16.7%)
 CPOs are burned out / overworked  14 (15.6%)
 Can not meet needs or Correctional Plans of offenders  12 (13.3%)
 Can not address recidivism rates or risk  12 (13.3%)
 Programs will not be offered  8 (8.9%)
 Offenders experience frustration / increased incidents / decreased motivation  8 (8.9%)
 Offenders are not cascaded down to lower security  4 (4.4%)
 Language barriers / programs not offered in language of choice  3 (3.3%)
 Delay in delivery of programs  2 (2.2%)
 Programs are of lower quality  1 (1.1%)
 Lack of training of current and new CPOs  1 (1.1%)
 Increased outside contracting  1 (1.1%)

 

Table. I13 Staff suggestions to improve the efficiency of CPO training
 Suggestions  Staff Respondents N (%)
 Provide an orientation / generic skills training first  27 (23.9%)
 Offer ongoing training / refreshers / mentoring  18 (15.9%)
 Longer training sessions  16 (14.2%)
 More frequent training  11 (9.7%)
 More training locations (nationally)  9 (8.0%)
 More training on dealing with offenders / responsivity issues  9 (8.0%)
 More time on content  7 (6.2%)
 More time on delivery / motivation  7 (6.2%)
 Hire qualified / certified people  7 (6.2%)
 More resources  5 (4.4%)
 More job shadowing / sitting in on program  5 (4.4%)
 Implement monitoring  3 (2.7%)
 Training via computer / video conference  3 (2.7%)
 More training on report writing  3 (2.7%)
 Establish national training schedules  3 (2.7%)
 More time pre-training for reading course materials  3 (2.7%)
 More culturally specific training  2 (1.8%)
 More time on group dynamics / mental health  2 (1.8%)
 Streamline testing aspect  2 (1.8%)
 Train staff on multiple programs  2 (1.8%)
 More contact between CPOs and trainers  1 (<1%)
 Less time on program delivery  1 (<1%)
 Less presentations during training  1 (<1%)
 More training for trainers  1 (<1%)
 More role playing  1 (<1%)
 Smaller group sizes for training  1 (<1%)
Note: Several respondents (N = 12, 10.6%) provided suggestions unrelated to the question.

 

Table I14. Recommendations for improving the OIA as suggested by staff members
 Recommendations:  Staff Respondents N (%)
 More training/certification for intake staff  18 (11.1%)
 Have CPOs working at intake  18 (11.1%)
 Offer programs to offenders during OIA  14 (8.6%)
 More resources needed  14 (8.6%)
 Need computer application with program availability database  12 (7.4%)
 Shorter assessment periods for short sentences  11 (6.8%)
 More psychological assessments/info  11 (6.8%)
 Increase communication between intake staff  10 (6.2%)
 Have more experienced staff administer the OIA  10 (6.2%)
 More one-on-one interviews  9 (5.6%)
 Reduce/eliminate the use of self-report  8 (4.9%)
 More cooperation/teamwork throughout OIA  8 (4.9%)
 Increase the time allotted to completed the OIA  8 (4.9%)
 Ensure all assessments completed before pen placed  8 (4.9%)
 Dedicated resources for Aboriginals at intake  8 (4.9%)
 Decrease the time taken to obtain community information  8 (4.9%)
 Centralize intake  8 (4.9%)
 Stop over-programming/address most pressing needs only  7 (4.3%)
 More communication between sites (intake/pens/community)  7 (4.3%)
 More information for offenders re purpose of programs  6 (3.7%)
 More clarity in program decision/placement  6 (3.7%)
 Increase accuracy of risk assessments  5 (3.1%)
 Decrease the # of instruments/assessments  5 (3.1%)
 More computer automation to decrease data entry duplication  4 (2.5%)
 Make OIA more applicable to women  4 (2.5%)
 Increase the accuracy/detail of assessments  4 (2.5%)
 Increase accuracy of community reports  4 (2.5%)
 Re-assess admission criteria for programs  3 (1.9%)
 More offender background info/histories  3 (1.9%)
 Use research-based assessments only  2 (1.2%)
 Update the entire OIA - make more current  2 (1.2%)
 OIA should be 12 hours/day (less downtime)  2 (1.2%)
 Need structured timelines for program completion  2 (1.2%)
 Implement a 2-tiered intake process  2 (1.2%)
 Do intake at remand centres  2 (1.2%)
 Adjust timing of OIA to prevent overwhelming offenders  2 (1.2%)
 More emphasis on inmates mother tongue  1 (<1%)
 Improve offender movement at intake  1 (<1%)
 Less use of Elders/more standard assessments for Aboriginals  1 (<1%)
 Decrease individual ability to over-ride assessment results  1 (<1%)
 Conduct supplementary assessments earlier in process  1 (<1%)
Note: Several respondents (4.3%, N = 7) provided answers unrelated to the question; these responses are not listed.

 

Table I15. Staff suggestions to increase administrative level support for correctional programs
 Suggestions  Staff Respondents N (%)
 Increase funding / staff  24 (34.8%)
 Increase number of CPOs  8 (11.6%)
 Improve communication between administration and frontline staff  5 (7.2%)
 Offer more programs in general  5 (7.2%)
 Increase clerical / data entry resources  4 (5.8%)
 Emphasize importance of Correctional Programs among staff  4 (5.8%)
 Consider offender movement/security when planning programs  4 (5.8%)
 Allow more time for report writing  2 (2.9%)
 Increase French programming for offenders  2 (2.9%)
 Allow more flexibility with resources / direct resources more efficiently  2 (2.9%)
 Increase Aboriginal-specific programs  1 (1.4%)
 Provide programs for offenders with special needs  1 (1.4%)
 Improve report editing / distribution process  1 (1.4%)
 Improve organization / location of program offices  1 (1.4%)
 Increase collaboration with psychologists for high intensity programs  1 (1.4%)
 Balance Correctional Programs with offender employment  1 (1.4%)
 Staff should encourage offenders to get involved in correctional programs  1 (1.4%)
 Staff positions more appropriately  1 (1.4%)
 Increase staff training / awareness of culturally sensitive issues  1 (1.4%)
 Increase administrative support for women offender programs  1 (1.4%)
 Do not call offenders out of group for administrative reasons  1 (1.4%)
 Provide training for program managers  1 (1.4%)
 Train staff on the use of electronic systems  1 (1.4%)
 Ensure electronic resources (e.g., Infonet) are updated  1 (1.4%)
 Increase availability of staff training in language of choice  1 (1.4%)
Note: Several respondents (13.0%, N = 9) provided suggestions that were not related to the question.

 

Table I16. Staff suggestions to increase management level support for correctional programs
 Suggestions  Staff Respondents N (%)
 Increase funding/staff/direct resources  35 (38.0%)
 Improve communication/availability between management and frontline staff  20 (21.7%)
 Increase staff awareness/emphasize importance of correctional programs among staff  15 (16.3%)
 Provide more programs/offer programs more often  6 (6.5%)
 Train program managers in specific program material  6 (6.5%)
 Provide staff with research/documentation on correctional programs  5 (5.4%)
 Offer sufficient training for staff  5 (5.4%)
 Clarify CPO duties / reduce non-program duties for CPOs  4 (4.3%)
 Increase involvement of Elders / give Elders more resources  3 (3.3%)
 Reduce lockdowns / consider offender movement issues  3 (3.3%)
 Ensure availability of backup managerial staff when RHQ management is not available  2 (2.2%)
 Improve balance between correctional programs and offender employment  2 (2.2%)
 Allow more time for report writing  1 (1.1%)
 Update relevant information / policies  1 (1.1%)
 Increase availability of French programs for offenders  1 (1.1%)
 Increase fairness of CPO performance evaluations  1 (1.1%)
 Reduce Parole Officer caseloads  1 (1.1%)
 Increase community programs  1 (1.1%)
 Do not call offenders out of group for administrative reasons  1 (1.1%)
 Reduce managerial staff and increase frontline staff  1 (1.1%)
 Streamline correctional programs for offenders with short sentences  1 (1.1%)
Note: Several respondents (9.8%, N = 9) provided suggestions that were not related to the question.

 

Table I17. Staff suggestions to increase NHQ support for correctional programs
 Suggestions:  Staff Respondents N (%)
 Improve communication/leadership between NHQ and frontline / regional staff  6 (23.1%)
 Increase funding / staff  6 (23.1%)
 Increase NHQ’s understanding of site operations  5 (19.2%)
 Provide research/documentation on correctional programs to managers  3 (11.5%)
 Update information & policies in a timely manner / provide guidelines  3 (11.5%)
 Update correctional programs to reflect changing offender profile  2 (7.7%)
 Provide stability in staffing management  2 (7.7%)
 Reduce lockdowns / consider offender movement issues  1 (3.8%)
 Offer core programs more often in institutions  1 (3.8%)
 Facilitate the accreditation of Aboriginal programs  1 (3.8%)
 Increase accountability of staff at site  1 (3.8%)
 Improve balance between correctional programs and offender employment  1 (3.8%)
Note: Several respondents (15.4%, N = 4) provided suggestions that were not related to the question.

 

Table I18. Suggestions for Best Practices in the Implementation of Correctional Programs
 Recommendations:  Staff Respondents N (%)
 More resources ($, staff, staff support)  35 (20.1%)
 More staff training/mentoring  33 (19.0%)
 More programs at all security levels  27 (15.5%)
 More flexibility in program delivery  24 (13.8%)
 More communication b/w staff/sites/NPB  22 (12.6%)
 More community support/ programs  20 (11.5%)
 Hire professional/certified staff/degree requirement  12 (6.9%)
 Move to modular program delivery  12 (6.9%)
 More accuracy in program placement/assessment  11 (6.3%)
 More set/continuous program schedules  10 (5.7%)
 More 1-on-1 programming/individual interaction  10 (5.7%)
 More team-oriented approach  10 (5.7%)
 More flexibility in program referral/placement  10 (5.7%)
 Decrease over-programming  7 (4.0%)
 More maintenance/booster programs  7 (4.0%)
 More ethno-cultural representation in programs  5 (2.9%)
 More use of psychologists/psychological services  5 (2.9%)
 More resources for those with short sentences  5 (2.9%)
 More consistency in programs (nationally)  5 (2.9%)
 Implement dual facilitator program delivery  4 (2.3%)
 More centralized services/intake  3 (1.7%)
 Educate offenders on correctional process/ programs  3 (1.7%)
 Evaluate programs to see what works  3 (1.7%)
 More Aboriginal staff  2 (1.1%)
 More variety in staff training methods  2 (1.1%)
 More focus on women offenders  2 (1.1%)
 More focus on integration  2 (1.1%)
 Create updated program matrix  2 (1.1%)
 More program delivery space  2 (1.1%)
 Increased efficiency of data management  1 (<1 %)
 Decrease program group sizes  1 (<1%)
 Implement formal tutoring program  1 (<1%)
 More justification in overriding program referrals  1 (<1%)
 More efficient/direct/accurate program reports  1 (<1%)
 More skills/vocational training  1 (<1%)
 Develop workload formula for CPOs  1 (<1%)
 Allot space in each program for lifers  1 (<1%)
Note: Several respondents (4.0%, N = 7) provided answers unrelated to the question; these responses are not listed.

 

Table I19. Trends in Dynamic Needs Ratings for Offender Sub-populations
 Needs Domain  Non-Aboriginal Male Offenders93  Aboriginal Male Offenders94  Women Offenders95
 Overall Dynamic Need  Proportion rated high need increased from 63% to 68%

 Decrease in moderate need rating (31% to 27%)

 Decrease in low need rating (6% to 4%)

 Proportion rated high need: First Nations 78%; Métis 73%; Inuit 89%

 Overall need rating; 57.3% some need, 65.1% considerable need96

 Proportion rated high need increased from 43% to 50%

 Relatively stable proportion of women rated as moderate need

 Decrease proportion rated as low need 19% to 12%)

 Employment/ Education  Considerable need rating decreased from 36% to 19%

 Some need rating decreased from 40% to 37%

 Rated some or considerable need: First Nations 70%; Métis 71%; Inuit 57%  Considerable need decreased from 24% to 15%

 Some need decreased from 53% to 41%

 Marital/Family  Considerable need rating decreased from 31% to 19%

 Some need rating decreased from 33% to 27%

 Rated some or considerable need: First Nations 60%; Métis 53%; Inuit 73%  Considerable need was relatively unchanged (29% to 27%)

 Some need rating decreased from 48% to 38%

 Associates  Considerable need rating decreased from 34% to 27%

 Some need rating was unchanged at 38%

 Rated some or considerable need: First Nations 65%; Métis 70%; Inuit 55%  20% assessed as having treatment needs in this area

 Considerable need rating was unchanged at 18%

Some need rating decreased from 58% to 42%

 Substance Abuse  Decrease in considerable risk rating (55% to 52%)

 Some need rating was unchanged at 21%

 Rated some or considerable need: First Nations 94%; Métis 91%; Inuit 92%  40-60% were assessed as having significant substance abuse problems

 Significant increase in considerable need rating from 42% to 56%

 Some need rating decreased from 22% to 15%

 Community Functioning  Considerable need rating decreased from 22% to 12%

 Some need rating decreased from 44% to 31%

 Rated some or considerable need: First Nations 45%; Métis 47%; Inuit 44%  Considerable need rating was relatively unchanged (8% to 7%)

 Some need rating decreased from 60% to 31%

 Personal/Emotional  Increase in considerable need rating from 68% to 72%

 Some need rating decreased from 25% to 21%

 Rated some or considerable need: First Nations 96%; Métis 95%; Inuit 99%  60% assessed as having considerable need in this area

 Increase in considerable need rating (43% to 63%)

 Decrease in some need rating (48% to 26%)

 Attitude  Considerable need rating decreased from 39% to 36%

 Some need rating was relatively unchanged (29% to 28%)

 Rated some or considerable need: First Nations 53%; Métis 56%; Inuit 53%  14% assessed as having attitude problems

 Considerable need rating was relatively unchanged (12% to 14%)

 Increase in some need rating from 18% to 26%

Overall dynamic needs ratings

An offender’s overall dynamic factor rating has implications for the level of service delivery (the more needs an offender has, the more programs he or she might require) and program intensity levels (increasing demand for high intensity programs). Increases in the proportion of high needs offenders suggests the continued relevancy of multiple programs targeting a range of needs domains, as well as access to higher intensity programs.

Employment/Education

Educational needs are important in determining whether an offender is eligible to participate in correctional programs. As trends in this domain suggest, a substantial proportion of offenders likely require some form of educational intervention prior to participating in correctional programs.

Marital/Family

Marital/family needs are targeted by a number of correctional programs, including FVPPs, WOSAP, the Tupiq Program for Inuit Offenders, SOAW, AOSAP, Aboriginal Basic Healing, Circles of Change, and Community Maintenance Programs (including the Women Offenders Relapse Prevention/Maintenance Program). Trends suggest a general decrease in marital/family needs among federal offenders. However, needs in this area continue to be a contributing factor for a substantial group of offenders, and suggest the continued relevance of correctional programs targeting this domain.

Associates

Trends seen in the associates domain suggest the continued relevance of programs targeting this criminogenic factor, programs such as ISOYW, AOSAP, and Counter-Point assist in this process. Although high needs ratings in this domain appear to be decreasing in recent years, it is apparent that a significant number of federal offenders still require interventions targeting antisocial associates.

Substance Abuse

It is apparent that a substantial proportion of the federal offender population has substance abuse needs, and requires correctional programs targeting this area. Specifically, the large proportions of offenders rated as having “considerable” substance abuse needs suggests the continued relevance of high intensity substance abuse programs. Substance abuse needs are targeted by SAPs, and this domain becomes relevant both in terms of program referral criteria and targeted behaviour.

Community Functioning

Community functioning is a domain targeted by a number of programs, including VOPs, SAPs, SOPs, community integration, and community maintenance programs. Once again, although there may be decreasing needs in this domain, it is apparent that a significant number of federal offenders still require interventions targeting community functioning.

Personal/Emotional

The substantial proportion of offenders experiencing needs in this domain suggests the continued relevance of the wide range of programs targeting personal/emotional needs. These include FVPPs, VOPs, SOPs, SAPs, the Tupiq Program for Inuit Offenders, SOAW, ISOYW, Aboriginal Basic Healing, Circles of Change, ABC, R & R, A & EM, Cognitive Skills Maintenance, and Community Integration and Maintenance Programs.

Attitude

General attitude is a domain targeted by a number of correctional programs, including FVPPs, VOPs, SOPs, SAPs, ISOYW, SOAW, Aboriginal Basic Healing, Circles of Change, ABC, A & EM, Counter-Point, and Community Maintenance Programs. Large numbers of offenders continue to exhibit need in this area.

Table I20. Stakeholder Positions Represented by Staff Interview Respondents
 Position Title  N  %
 Parole Officer  73  30.3
 CPO/Program Facilitator  56  23.2
 Program Manager  20  8.3
 Regional Program Manager  17  7.1
 Assistant Warden, Intervention  13  5.4
 Manager, Assessment and Intervention  9  3.7
 Psychologist/Psychotherapist  8  3.3
 Parole Officer Supervisor  7  2.9
 Elder/Spiritual Advisor  7  2.9
 Aboriginal Liaison Officer  5  2.1
 Behavioural Counsellor  4  1.7
 Program Coordinator  3  1.2
 Regional Administrator, Correctional Programs  3  1.2
 Assistant Deputy Commissioner  3  1.2
 Case Management Officer  2  0.8
 Clinical Director/Clinical Team Leader  2  0.8
 Inmate Pay/Employment Officer  2  0.8
 Case Management Supervisor  1  0.4
 Correctional Manager, Psychiatric Hospital  1  0.4
 Personal Development Clerk  1  0.4
 Manager, Intensive Intervention Strategy  1  0.4
 Program Plan Assistant  1  0.4
 Guidance Counsellor  1  0.4
 Executive Director  1  0.4
 Unspecified  1  0.4

 

Appendix J: Summary of Data

 Effectiveness Indicator
 Program: VPP
 Intent to Treat  Comparison
 Institutional Outcomes  N = 502  N = 833
    Change in yearly rate of violent institutional incidents (mean)  -0.025 ns  -0.142**
    Change in yearly rate of non-violent institutional incidents (mean)  0.245 ns  -0.329 ns
    Likelihood of discretionary release (Odds Ratio)  3.054***  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 300, X = 1168.2  N = 464, X = 1106.6
    Difference between time at risk in community  T-statistic = -1.46 ns
    Likelihood of any readmission  0.712***  1
    Likelihood of readmission with new offence  0.588***  1
    Likelihood of readmission with new violent offence  0.483***  1
    Community outcomes where approximated referral criteria met  N = 255, X = 1178.8  N = 330, X = 1112.8
    Difference between time at risk in community  T-statistic = -1.44 ns
    Likelihood of any readmission  0.634***  1
    Likelihood of readmission with new offence  0.498***  1
    Likelihood of readmission with new violent offence  0.568*  1
    Completions  N = 233, X = 1188.7  N = 464, X = 1106.6
    Difference between time at risk in community  T-statistic = -1.8*
    Likelihood of any readmission  0.690***  1
    Likelihood of readmission with new offence  0.547***  1
    Likelihood of readmission with new violent offence  0.458**  1
    Drop-Outs  N = 67, X = 1096.9  N = 464, X = 1106.6
    Difference between time at risk in community  T-statistic = 0.14 ns
    Likelihood of any readmission  0.778 ns  1
    Likelihood of readmission with new offence  0.710 ns  1
    Likelihood of readmission with new violent offence  0.541 ns  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)  Intent to Treat  Comparison
 Institutional Outcomes  N = 198  N = 221
    Change in yearly rate of violent institutional incidents (mean)  -0.058 ns  -0.007 ns
    Change in yearly rate of non-violent institutional incidents (mean)  0.089 ns  0.460 ns
    Likelihood of discretionary release (Odds Ratio)  5.580***  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 109, X = 1306  N = 104, X = 1198.9
    Difference between time at risk in community  T-statistic = -1.24 ns
    Likelihood of any readmission  0.824 ns  1
    Likelihood of readmission with new offence  0.697 ns  1
    Likelihood of readmission with new violent offence  0.373*  1
    Community outcomes where approximated referral criteria met  N = 85, X = 1340.7  N = 80, X = 1226.7
    Difference between time at risk in community  T-statistic = -1.14 ns
    Likelihood of any readmission  0.658 ns  1
    Likelihood of readmission with new offence  0.375*  1
    Likelihood of readmission with new violent offence  0.893 ns  1
    Completions  N = 90, X = 1288.2  N = 104, X = 1198.9
    Difference between time at risk in community  T-statistic = -0.98 ns
    Likelihood of any readmission  0.765 ns  1
    Likelihood of readmission with new offence  0.624*  1
    Likelihood of readmission with new violent offence  0.459 ns  1
    Drop-Outs  N = 19, X = 1390.5  N = 104, X = 1198.9
    Difference between time at risk in community  T-statistic = -1.32 ns
    Likelihood of any readmission  1.081 ns  1
    Likelihood of readmission with new offence  1.017 ns  1
    Likelihood of readmission with new violent offence  0 ns  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Non-Aboriginal)
 Program: ABC
 Intent to Treat  Comparison
 Institutional Outcomes  N = 242  N = 238
    Change in yearly rate of violent institutional incidents (mean)  -0.054 ns  -0.027 ns
    Change in yearly rate of non-violent institutional incidents (mean)  -1.331 ns  -1.015 ns
    Likelihood of discretionary release (Odds Ratio)  2.266**  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 186, X = 1589.2  N = 120, X = 1357.3
    Difference between time at risk in community  T-statistic = -2.89***
    Likelihood of any readmission  0.893 ns  1
    Likelihood of readmission with new offence  0.565**  1
    Likelihood of readmission with new violent offence  0.467*  1
    Community outcomes where approximated referral criteria met  N = 75, X = 1886.5  N = 44, X = 1568
    Difference between time at risk in community  T-statistic = -2.36**
    Likelihood of any readmission  0.607**  1
    Likelihood of readmission with new offence  0.364***  1
    Likelihood of readmission with new violent offence  0.360*  1
    Completions  N = 176, X = 1627.8  N = 120, X = 1357.3
    Difference between time at risk in community  T-statistic = -3.34***
    Likelihood of any readmission  0.848 ns  1
    Likelihood of readmission with new offence  0.506***  1
    Likelihood of readmission with new violent offence  0.425*  1
    Drop-Outs  N = 10, X = 910.9  N = 120, X = 1357.3
    Difference between time at risk in community  T-statistic = 2.19**
    Likelihood of any readmission  1.909*  1
    Likelihood of readmission with new offence  2.057 ns  1
    Likelihood of readmission with new violent offence  1.967 ns  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 Program: ABC
 Intent to Treat  Comparison
 Institutional Outcomes  N = 168  N = 74
    Change in yearly rate of violent institutional incidents (mean)  -0.091 ns  -0.133*
    Change in yearly rate of non-violent institutional incidents (mean)  -0.486 ns  -0.391 ns
    Likelihood of discretionary release (Odds Ratio)  2.785 ns  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 129, X = 1757.5  N = 40, X = 1439.1
    Difference between time at risk in community  T-statistic = -2.47**
    Likelihood of any readmission  0.639**  1
    Likelihood of readmission with new offence  0.714 ns  1
    Likelihood of readmission with new violent offence  0.431*  1
    Community outcomes where approximated referral criteria met  N = 93, X = 1753.1  N = 29, X = 1318.6
    Difference between time at risk in community  T-statistic = -2.95***
    Likelihood of any readmission  0.506***  1
    Likelihood of readmission with new offence  0.543 ns  1
    Likelihood of readmission with new violent offence  0.171***  1
    Completions  N = 40, X = 1439.1  N = 118, X = 1797.6
    Difference between time at risk in community  T-statistic = -2.75***
    Likelihood of any readmission  0.653**  1
    Likelihood of readmission with new offence  0.729 ns  1
    Likelihood of readmission with new violent offence  0.468 ns  1
    Drop-Outs  N = 11, X = 1327.8  N = 40, X = 1439.1
    Difference between time at risk in community  T-statistic = 0.52 ns
    Likelihood of any readmission  0.408*  1
    Likelihood of readmission with new offence  0.690 ns  1
    Likelihood of readmission with new violent offence  0 ns  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Non-Aboriginal)
 Program: ISOYW
 Intent to Treat  Comparison
 Institutional Outcomes  N = 24  N = 14
    Change in yearly rate of violent institutional incidents (mean)  0.0098 ns  -0.054 ns
    Change in yearly rate of non-violent institutional incidents (mean)  -0.273 ns  -2.115*
    Likelihood of discretionary release (Odds Ratio)  0.873 ns  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 17, X = 1204.6  N = 10, X = 1617.7
    Difference between time at risk in community  T-statistic = 1.66 ns
    Likelihood of any readmission  0.827 ns  1
    Likelihood of readmission with new offence  0.470 ns  1
    Likelihood of readmission with new violent offence  --  1
    Community outcomes where approximated referral criteria met  N = 6, X = 1172  N = 4, X = 1826.5
    Difference between time at risk in community  T-statistic = 1.17 ns
    Likelihood of any readmission  7.972 ns  1
    Likelihood of readmission with new offence  0 ns  1
    Likelihood of readmission with new violent offence  --  --
    Completions
    Likelihood of any readmission  --  --
    Likelihood of readmission with new offence  --  --
    Likelihood of readmission with new violent offence  --  --
    Drop-Outs
    Likelihood of any readmission  --  --
    Likelihood of readmission with new offence  --  --
    Likelihood of readmission with new violent offence  --  --

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 Program: ISOYW
 Intent to Treat  Comparison
 Institutional Outcomes  N = 431  N = 239
    Change in yearly rate of violent institutional incidents (mean)  -0.002 ns  -0.012 ns
    Change in yearly rate of non-violent institutional incidents (mean)  -0.316*  -0.095 ns
    Likelihood of discretionary release (Odds Ratio)  2.372***  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 267, X = 1237.1  N = 147, X = 1291.4
    Difference between time at risk in community  T-statistic = 0.90 ns
    Likelihood of any readmission  0.809*  1
    Likelihood of readmission with new offence  0.893 ns  1
    Likelihood of readmission with new violent offence  0.795 ns  1
    Community outcomes where approximated referral criteria met  N = 159, X = 1239.8  N = 78, X = 1319.8
    Difference between time at risk in community  T-statistic = 0.93 ns
    Likelihood of any readmission  0.779 ns  1
    Likelihood of readmission with new offence  0.847 ns  1
    Likelihood of readmission with new violent offence  0.785 ns  1
    Completions  N = 249, X = 1262.2  N = 147, X = 1291.4
    Difference between time at risk in community  T-statistic = 0.47 ns
    Likelihood of any readmission  0.764**  1
    Likelihood of readmission with new offence  0.840 ns  1
    Likelihood of readmission with new violent offence  0.773 ns  1
    Completions  N = 18, X = 890.17  N = 147, X = 1291.4
    Difference between time at risk in community  T-statistic = 4.86***
    Likelihood of any readmission  1.833**  1
    Likelihood of readmission with new offence  1.931 ns  1
    Likelihood of readmission with new violent offence  1.217 ns  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Non-Aboriginal)
 Program: HIFVPP
 Intent to Treat  Comparison
 Institutional Outcomes  N = 487  N = 476
    Change in yearly rate of violent institutional incidents (mean)  0.190**  0.062 ns
    Change in yearly rate of non-violent institutional incidents (mean)  0.414**  0.217 ns
    Likelihood of discretionary release (Odds Ratio)  3.310***  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 294, X = 1230.7  N = 151, X = 1400.5
    Difference between time at risk in community  T-statistic = 2.64***
    Likelihood of any readmission  0.726**  1
    Likelihood of readmission with new offence  0.793 ns  1
    Likelihood of readmission with new violent offence  1.669 ns  1
    Community outcomes where approximated referral criteria met  N = 115, X = 772.33  N = 43, X = 802.23
    Difference between time at risk in community  T-statistic = 0.56 ns
    Likelihood of any readmission  0.814 ns  1
    Likelihood of readmission with new offence  0.679 ns  1
    Likelihood of readmission with new violent offence  0.539 ns  1
    Completions  N = 264, X = 1249.6  N = 151, X = 1400.5
    Difference between time at risk in community  T-statistic = 2.28**
    Likelihood of any readmission  0.716**  1
    Likelihood of readmission with new offence  0.769 ns  1
    Likelihood of readmission with new violent offence  1.786 ns  1
    Completions  N = 30, X = 1063.9  N = 151, X = 1400.5
    Difference between time at risk in community  T-statistic = 2.60**
    Likelihood of any readmission  0.793 ns  1
    Likelihood of readmission with new offence  0.762 ns  1
    Likelihood of readmission with new violent offence  0 ns  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 Program: HIFVPP
 Intent to Treat  Comparison
 Institutional Outcomes  N = 148  N = 107
    Change in yearly rate of violent institutional incidents (mean)  -0.003 ns  0.050 ns
    Change in yearly rate of non-violent institutional incidents (mean)  -0.012 ns  0.140 ns
    Likelihood of discretionary release (Odds Ratio)  1.187 ns  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 108, X = 1210.1  N = 43, X = 1384.4
    Difference between time at risk in community  T-statistic = 1.78***
    Likelihood of any readmission  0.815 ns  1
    Likelihood of readmission with new offence  0.977 ns  1
    Likelihood of readmission with new violent offence  0.740 ns  1
    Community outcomes where approximated referral criteria met  N = 40, X = 817.35  N = 16, X = 789.56
    Difference between time at risk in community  T-statistic = -0.30 ns
    Likelihood of any readmission  0.826 ns  1
    Likelihood of readmission with new offence  2.812 ns  1
    Likelihood of readmission with new violent offence  3.757 ns  1
    Completions  N = 99, X = 1203.4  N = 43, X = 1384.4
    Difference between time at risk in community  T-statistic = 1.82*
    Likelihood of any readmission  0.739 ns  1
    Likelihood of readmission with new offence  0.796 ns  1
    Likelihood of readmission with new violent offence  0.545 ns  1
    Completions  N = 9, X = 1283.2  N = 43, X = 1384.4
    Difference between time at risk in community  T-statistic = 0.46 ns
    Likelihood of any readmission  1.058 ns  1
    Likelihood of readmission with new offence  1.090 ns  1
    Likelihood of readmission with new violent offence  7.010 ns  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Non-Aboriginal)
 Program: MIFVPP
 Intent to Treat  Comparison
 Institutional Outcomes  N = 758  N = 911
    Change in yearly rate of violent institutional incidents (mean)  0.04 ns  -0.004 ns
    Change in yearly rate of non-violent institutional incidents (mean)  0.254*  0.125 ns
    Likelihood of discretionary release (Odds Ratio)  1.70***  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 450, X = 1108.2  N = 530, X = 1207.8
    Difference between time at risk in community  T-statistic = 3.05***
    Likelihood of any readmission  0.745***  1
    Likelihood of readmission with new offence  0.637***  1
    Likelihood of readmission with new violent offence  0.426**  1
    Community outcomes where approximated referral criteria met  N = 42, X = 922.52  N = 49, X = 870.43
    Difference between time at risk in community  T-statistic = -0.72 ns
    Likelihood of any readmission  0.642*  1
    Likelihood of readmission with new offence  0.555 ns  1
    Likelihood of readmission with new violent offence  0.615 ns  1
    Completions  N = 166, X = 821.77  N = 152, X = 826.36
    Difference between time at risk in community  T-statistic = 0.13
    Likelihood of any readmission  0.707**  1
    Likelihood of readmission with new offence  0.582**  1
    Likelihood of readmission with new violent offence  0.432 ns  1
    Completions  N = 49, X = 1021.1  N = 530, X = 1207.8
    Difference between time at risk in community  T-statistic = 3.12***
    Likelihood of any readmission  0.777 ns  1
    Likelihood of readmission with new offence  0.677 ns  1
    Likelihood of readmission with new violent offence  0.313 ns  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 Program: MIFVPP
 Intent to Treat  Comparison
 Institutional Outcomes  N = 234  N = 272
    Change in yearly rate of violent institutional incidents (mean)  0.074 ns  0.010 ns
    Change in yearly rate of non-violent institutional incidents (mean)  0.385 ns  -0.507**
    Likelihood of discretionary release (Odds Ratio)  1.012 ns  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 150, X = 1101.5  N = 150, X = 1150.9
    Difference between time at risk in community  T-statistic = 0.87 ns
    Likelihood of any readmission  0.868 ns  1
    Likelihood of readmission with new offence  0.962 ns  1
    Likelihood of readmission with new violent offence  1.153 ns  1
    Community outcomes where approximated referral criteria met  N = 22, X = 812  N = 21, X = 929.45
    Difference between time at risk in community  T-statistic = -1.16 ns
    Likelihood of any readmission  0.560 ns  1
    Likelihood of readmission with new offence  1.052 ns  1
    Likelihood of readmission with new violent offence  0.820 ns  1
    Completions  N = 136, X = 1100.8  N = 150, X = 1150.9
    Difference between time at risk in community  T-statistic = 0.86 ns
    Likelihood of any readmission  0.820 ns  1
    Likelihood of readmission with new offence  0.978 ns  1
    Likelihood of readmission with new violent offence  1.248 ns  1
    Completions  N = 14, X = 1108.3  N = 150, X = 1150.9
    Difference between time at risk in community  T-statistic = 0.29 ns
    Likelihood of any readmission  1.618 ns  1
    Likelihood of readmission with new offence  0.934 ns  1
    Likelihood of readmission with new violent offence  0 ns  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 Program: HI-NAFVPP
 Intent to Treat  Comparison
 Institutional Outcomes  N = 60  N = 19
    Change in yearly rate of violent institutional incidents (mean)  -0.072 ns  -0.239 ns
    Change in yearly rate of non-violent institutional incidents (mean)  -0.115 ns  -0.356 ns
    Likelihood of discretionary release (Odds Ratio)  <0.001 ns  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 15, X = 770.2  N = 1, X = 792
    Difference between time at risk in community  T-statistic = 0.07 ns
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  0.664 ns  1
    Likelihood of readmission with new violent offence  0.903 ns  1
    Community outcomes where approximated referral criteria met  N = 11, X = 684  N = 1, X = 792
    Difference between time at risk in community  T-statistic = 0.32 ns
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  0,918 ns  1
    Likelihood of readmission with new violent offence  0.920 ns  1
    Completions  N = 14, X = 790.57  N = 1, X = 792
    Difference between time at risk in community  T-statistic = 0.00 ns
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  0.640 ns  1
    Likelihood of readmission with new violent offence  0.906 ns  1
    Completions  N = 1, X = 485  N = 1, X = 792
    Difference between time at risk in community  T-statistic = --
    Likelihood of any readmission  2.398 ns  1
    Likelihood of readmission with new offence  2.398 ns  1
    Likelihood of readmission with new violent offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Non-Aboriginal)
 Program: NSAP-High
 Intent to Treat  Comparison
 Institutional Outcomes  N = 611  N = 668
    Change in yearly rate of substance-related institutional incidents (mean)  0.076 ns  0.093 ns
    Change in yearly rate of non-substance-related institutional incidents (mean)  0.940***  0.207 ns
    Likelihood of discretionary release (Odds Ratio)  4.51***  1
    Community Outcomes (ITT): Sample Size (N), Mean time at risk (X)  N = 221, X = 567.45  N = 219, X = 846.1
    Difference between time at risk in community  T-statistic = 9.0***
    Likelihood of any readmission  0.849 ns  1
    Likelihood of readmission with new offence  0.552***  1
    Likelihood of readmission with new violent offence  0.368***  1
    Community outcomes where approximated referral criteria met (ITT)  N = 36, X = 557.25  N = 26, X = 805.92
    Difference between time at risk in community  T-statistic = 3.27***
    Likelihood of any readmission  1.369 ns  1
    Likelihood of readmission with new offence  0.744 ns  1
    Likelihood of readmission with new violent offence  0.100 ns  1
    Completions  N = 181, X = 568.767  N = 219, X = 846.1
    Difference between time at risk in community  T-statistic = 8.86***
    Likelihood of any readmission  0.761**  1
    Likelihood of readmission with new offence  0.530***  1
    Likelihood of readmission with new violent offence  0.250***  1
    Completions  N = 40, X = 561.5  N = 219, X = 846.1
    Difference between time at risk in community  T-statistic = 7.48***
    Likelihood of any readmission  1.460*  1
    Likelihood of readmission with new offence  0.730 ns  1
    Likelihood of readmission with new violent offence  1.290 ns  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 Program: NSAP-High
 Intent to Treat  Comparison
 Institutional Outcomes  N = 147  N = 195
    Change in yearly rate of substance-related institutional incidents (mean)  0.172 ns  -0.151 ns
    Change in yearly rate of non-substance-related institutional incidents (mean)  -0.153 ns  0.382 ns
    Likelihood of discretionary release (Odds Ratio)  7.151**  1
    Community Outcomes (ITT): Sample Size (N), Mean time at risk (X)  N = 47, X = 542.6  N = 61, X = 727.4
    Difference between time at risk in community  T-statistic = 4.13***
    Likelihood of any readmission  0.631*  1
    Likelihood of readmission with new offence  0.609 ns  1
    Likelihood of readmission with new violent offence  0.181 ns  1
    Community outcomes where approximated referral criteria met (ITT)  N = 6, X = 517.5  N = 5, X = 675.8
    Difference between time at risk in community  T-statistic = 1.24 ns
    Likelihood of any readmission  0 ns  1
    Likelihood of readmission with new offence  0 ns  1
    Likelihood of readmission with new violent offence  --  --
    Completions  N = 37, X = 535.14  N = 61, X = 727.48
    Difference between time at risk in community  T-statistic = 4.27***
    Likelihood of any readmission  0.527**  1
    Likelihood of readmission with new offence  0.518 ns  1
    Likelihood of readmission with new violent offence  0.196 ns  1
    Completions  N = 10, X = 570.2  N = 61, X = 727.48
    Difference between time at risk in community  T-statistic = 2.56**
    Likelihood of any readmission  1.178 ns  1
    Likelihood of readmission with new offence  2.192 ns  1
    Likelihood of readmission with new violent offence  0 ns  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Non-Aboriginal)
 Program: NSAP-Moderate
 Intent to Treat  Comparison
 Institutional Outcomes  N = 2.428  N = 1.079
    Change in yearly rate of substance-related institutional incidents (mean)  0.104***  -0.031 ns
    Change in yearly rate of non-substance-related institutional incidents (mean)  0.307***  -0.261 ns
    Likelihood of discretionary release (Odds Ratio)  1.434***  1
    Community Outcomes (ITT): Sample Size (N), Mean time at risk (X)  N = 1,004, X = 629.78  N = 536, X = 887.89
    Difference between time at risk in community  T-statistic = 13.07***
    Likelihood of any readmission  0,819***  1
    Likelihood of readmission with new offence  0,735***  1
    Likelihood of readmission with new violent offence  0,543**  1
    Community outcomes where approximated referral criteria met (ITT)  N = 627, X = 629.16  N = 286, X = 894.91
    Difference between time at risk in community  T-statistic = 13.60***
    Likelihood of any readmission  0.832**  1
    Likelihood of readmission with new offence  0.723**  1
    Likelihood of readmission with new violent offence  0.330***  1
    Completions  N = 924, X = 626.37  N = 536, X = 887.89
    Difference between time at risk in community  T-statistic = 13.21***
    Likelihood of any readmission  0.810***  1
    Likelihood of readmission with new offence  0.714***  1
    Likelihood of readmission with new violent offence  0.481***  1
    Completions  N = 80, X = 669.21  N = 536, X = 887.89
    Difference between time at risk in community  T-statistic = 7.84***
    Likelihood of any readmission  0.891 ns  1
    Likelihood of readmission with new offence  0.953 ns  1
    Likelihood of readmission with new violent offence  1.094 ns  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 Program: NSAP-Moderate
 Intent to Treat  Comparison
 Institutional Outcomes  N = 722  N = 362
    Change in yearly rate of substance-related institutional incidents (mean)  -0.015 ns  0.071 ns
    Change in yearly rate of non-substance-related institutional incidents (mean)  0.161 ns  -0.457*
    Likelihood of discretionary release (Odds Ratio)  1.368 ns  1
    Community Outcomes (ITT): Sample Size (N), Mean time at risk (X)  N = 236, X = 637.81  N = 148, X = 830.93
    Difference between time at risk in community  T-statistic = 5.52***
    Likelihood of any readmission  0.847 ns  1
    Likelihood of readmission with new offence  0.692*  1
    Likelihood of readmission with new violent offence  0.593 ns  1
    Community outcomes where approximated referral criteria met (ITT)  N = 167, X = 632.86  N = 82, X = 826.21
    Difference between time at risk in community  T-statistic = 4.67***
    Likelihood of any readmission  0.943 ns  1
    Likelihood of readmission with new offence  0.708 ns  1
    Likelihood of readmission with new violent offence  0.417 ns  1
    Completions  N = 219, X = 637.01  N = 148, X = 830.93
    Difference between time at risk in community  T-statistic = 5.52***
    Likelihood of any readmission  0.848 ns  1
    Likelihood of readmission with new offence  0.723*  1
    Likelihood of readmission with new violent offence  0.701 ns  1
    Completions  N = 17, X = 648.12  N = 148, X = 830.93
    Difference between time at risk in community  T-statistic = 3.36***
    Likelihood of any readmission  0.859  1
    Likelihood of readmission with new offence  0.279*  1
    Likelihood of readmission with new violent offence  0 ns  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator
 Program: AOSAP
 Intent to Treat  Comparison
 Institutional Outcomes  N = 173  N = 93
    Change in yearly rate of substance-related institutional incidents (mean)  0.030 ns  0.032 ns
    Change in yearly rate of non-substance-related institutional incidents (mean)  0.375 ns  0.478 ns
    Likelihood of discretionary release (Odds Ratio)  2.21 ns  1
    Community Outcomes where approximated referral criteria met: Sample Size     (N), Mean time at risk (X)  N = 38, X = 575.18  N = 38, X = 722.5
    Difference between time at risk in community  T-statistic = 2.99***
    Likelihood of any readmission  0.512**  1
    Likelihood of readmission with new offence  0.493 ns  1
    Likelihood of readmission with new violent offence  0.382 ns  1
    Completions  N = 31, X = 553.77  N = 38, X = 722.5
    Difference between time at risk in community  T-statistic = 3.22***
    Likelihood of any readmission  0.510**  1
    Likelihood of readmission with new offence  0.532 ns  1
    Likelihood of readmission with new violent offence  0.459 ns  1
    Completions  N = 7, X = 670  N = 38, X = 722.5
    Difference between time at risk in community  T-statistic = 0.50 ns
    Likelihood of any readmission  0.786 ns  1
    Likelihood of readmission with new offence  0.554 ns  1
    Likelihood of readmission with new violent offence  0 ns  1
    Community outcomes where approximated referral criteria met (Moderate and     High)  N = 26, X = 580.35  N = 22, X = 724.45
    Difference between time at risk in community  T-statistic = 2.80***
    Likelihood of any readmission  0.553 ns  1
    Likelihood of readmission with new offence  0.746 ns  1
    Likelihood of readmission with new violent offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Non-Aboriginal)
 Program: NSAP-Maintenance (Male)
 Intent to Treat  Comparison
 Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 3,353, X = 1424.4  N = 1,886, X = 1330.6
    Difference between time at risk in community  T-statistic = -4.07***
    Likelihood of any readmission  0.781***  1
    Likelihood of readmission with new offence  0.731***  1
    Likelihood of readmission with new violent offence  0.971 ns  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 Program: NSAP-Maintenance (Male)
 Intent to Treat  Comparison
 Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 742, X = 1453.3  N = 637, X = 1240.3
    Difference between time at risk in community  T-statistic = -5.88***
    Likelihood of any readmission  0.758***  1
    Likelihood of readmission with new offence  0.701***  1
    Likelihood of readmission with new violent offence  0.780 ns  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator
 Program: NSAP-Maintenance (Women)
 Intent to Treat  Comparison
 Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 271, X = 1053.2  N = 115, X = 1088.1
    Difference between time at risk in community  T-statistic = 0.62 ns
    Likelihood of any readmission  0.618***  1
    Likelihood of readmission with new offence  0.779 ns  1
    Likelihood of readmission with new violent offence  0.502 ns  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Non-Aboriginal)
 Program: HI-SOP (Non-Child) (ALL)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 204  N = 201
    Change in yearly rate of violent institutional incidents (mean)  0.032 ns  0.044 ns
    Change in yearly rate of non-violent institutional incidents (mean)  -0,019 ns  0.304 ns
    Likelihood of discretionary release (Odds Ratio)  3.661 ns  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 65, X = 1755.3  N = 35, X = 1502.5
    Difference between time at risk in community  T-statistic = -1.56 ns
    Likelihood of any readmission  0.752 ns  1
    Likelihood of readmission with new offence  0.931 ns  1
    Likelihood of readmission with new violent offence  0.593 ns  1
    Likelihood of new sex offence  0.289 ns  1
    Community outcomes where approximated referral criteria met  N = 29, X = 1611.2  N = 10, X = 1430.1
    Difference between time at risk in community  T-statistic = -0.63 ns
    Likelihood of any readmission  0.449 ns  1
    Likelihood of readmission with new offence  0.092 ns  1
    Likelihood of readmission with new violent offence  0.079 ns  1
    Likelihood of new sex offence  0.079 ns  1
    Completions  N = 60, X = 1828.1  N = 35, X = 1502.5
    Difference between time at risk in community  T-statistic = -2.02 ns
    Likelihood of any readmission  0.725 ns  1
    Likelihood of readmission with new offence  0.951 ns  1
    Likelihood of readmission with new violent offence  0.612 ns  1
    Likelihood of new sex offence  0.299 ns  1
    Completions  N = 5, X = 882.4  N = 35, X = 1502.5
    Difference between time at risk in community  T-statistic = 1.73*
    Likelihood of any readmission  0.979 ns  1
    Likelihood of readmission with new offence  0 ns  1
    Likelihood of readmission with new violent offence  0 ns  1
    Likelihood of new sex offence  0 ns  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Non-Aboriginal)
 Program: HI-SOP(Non-Child) (Regional)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 77  N = 48
    Change in yearly rate of violent institutional incidents (mean)  0.175 ns  -0.042 ns
    Change in yearly rate of non-violent institutional incidents (mean)  -0.446 ns  0.019 ns
    Likelihood of discretionary release (Odds Ratio)  --  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 24, X = 1719.4  N = 2, X = 987.5
    Difference between time at risk in community  T-statistic = -1.58 ns
    Likelihood of any readmission  0 ns  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Community outcomes where approximated referral criteria met  N = 16, X = 1723.9  N = 1, X = 1260
    Difference between time at risk in community  T-statistic = -0.63 ns
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = 23, X = 1733.7  N = 2, X = 987.5
    Difference between time at risk in community  T-statistic = -1.58 ns
    Likelihood of any readmission  0 ns  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = 1, X = 1390  N = 2, X = 987.5
    Difference between time at risk in community  T-statistic = -0.85 ns
    Likelihood of any readmission  0 ns  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Non-Aboriginal)
 Program: HI-SOP (Non-Child) (Local)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 127  N = 153
    Change in yearly rate of violent institutional incidents (mean)  -0.054 ns  0.070 ns
    Change in yearly rate of non-violent institutional incidents (mean)  -0.034 ns  0.393 ns
    Likelihood of discretionary release (Odds Ratio)  3.758 ns  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 41, X = 1776.3  N = 33, X = 1533.7
    Difference between time at risk in community  T-statistic = -1.26 ns
    Likelihood of any readmission  0.693 ns  1
    Likelihood of readmission with new offence  1.023 ns  1
    Likelihood of readmission with new violent offence  0.743 ns  1
    Likelihood of new sex offence  0.507 ns  1
    Community outcomes where approximated referral criteria met  N = 13, X = 1472.5  N = 9, X = 1449
    Difference between time at risk in community  T-statistic = -0.06 ns
    Likelihood of any readmission  0.636 ns  1
    Likelihood of readmission with new offence  0.224 ns  1
    Likelihood of readmission with new violent offence  0.224 ns  1
    Likelihood of new sex offence  0.224 ns  1
    Completions  N = 37, X = 1886.7  N = 33, X = 1533.7
    Difference between time at risk in community  T-statistic = -1.85*
    Likelihood of any readmission  0.705 ns  1
    Likelihood of readmission with new offence  1.071 ns  1
    Likelihood of readmission with new violent offence  0.791 ns  1
    Likelihood of new sex offence  0.533 ns  1
    Completions  N = 4, X = 755.5  N = 33, X = 1533.7
    Difference between time at risk in community  T-statistic = 1.93*
    Likelihood of any readmission  0.490 ns  1
    Likelihood of readmission with new offence  0 ns  1
    Likelihood of readmission with new violent offence  0 ns  1
    Likelihood of new sex offence  0 ns  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Non-Aboriginal)
 Program: HI-SOP (Child) (ALL)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 249  N = 165
    Change in yearly rate of violent institutional incidents (mean)  -0.027 ns  -0.098**
    Change in yearly rate of non-violent institutional incidents (mean)  -0.336***  0.038 ns
    Likelihood of discretionary release (Odds Ratio)  44.642 ns  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 76, X = 1612.8  N = 19, X = 1502.7
    Difference between time at risk in community  T-statistic = -0.56 ns
    Likelihood of any readmission  0.461 ns  1
    Likelihood of readmission with new offence  0.362 ns  1
    Likelihood of readmission with new violent offence  0.261 ns  1
    Likelihood of new sex offence  0.501 ns  1
    Community outcomes where approximated referral criteria met  N = 43, X = 1499.8  N = 6, X = 1755.8
    Difference between time at risk in community  T-statistic = 0.82 ns
    Likelihood of any readmission  0.476 ns  1
    Likelihood of readmission with new offence  0.360 ns  1
    Likelihood of readmission with new violent offence  0.154 ns  1
    Likelihood of new sex offence  --  1
    Completions  N = 75, X = 1623.8  N = 19, X = 1502.7
    Difference between time at risk in community  T-statistic = -0.61 ns
    Likelihood of any readmission  0.465 ns  1
    Likelihood of readmission with new offence  0.364 ns  1
    Likelihood of readmission with new violent offence  0.262 ns  1
    Likelihood of new sex offence  0.503 ns  1
    Completions  N = 1, X = 787  N = 19, X = 1502.7
    Difference between time at risk in community  T-statistic = 0.85 ns
    Likelihood of any readmission  0 ns  1
    Likelihood of readmission with new offence  0.189 ns  1
    Likelihood of readmission with new violent offence  0.189 ns  1
    Likelihood of new sex offence  0.980 ns  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Non-Aboriginal)
 Program: HI-SOP (Child) (Regional)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 93  N = 48
    Change in yearly rate of violent institutional incidents (mean)  0.013 ns  -0.15**
    Change in yearly rate of non-violent institutional incidents (mean)  -0.364***  0.662 ns
    Likelihood of discretionary release (Odds Ratio)  --  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = --, X = --  N = --, X = --
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Community outcomes where approximated referral criteria met  N = --, X = --  N = --, X = --
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = --, X = --  N = --, X = --
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = --, X = --  N = --, X = --
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Non-Aboriginal)
 Program: HI-SOP (Child) (Local)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 155  N = 117
    Change in yearly rate of violent institutional incidents (mean)
 -0.052 ns  -0.076 ns
    Change in yearly rate of non-violent institutional incidents (mean)  -0.312*  -0.218 ns
    Likelihood of discretionary release (Odds Ratio)  38.730 ns  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 56, X = 1640.1  N = 19, X = 1502.7
    Difference between time at risk in community  T-statistic = -0.66 ns
    Likelihood of any readmission  0.469 ns  1
    Likelihood of readmission with new offence  0.333 ns  1
    Likelihood of readmission with new violent offence  0.333 ns  1
    Likelihood of new sex offence  0.643 ns  1
    Community outcomes where approximated referral criteria met  N = 26, X = 1466.8  N = 6, X = 1755.8
    Difference between time at risk in community  T-statistic = 0.84 ns
    Likelihood of any readmission  0.426 ns  1
    Likelihood of readmission with new offence  0.468 ns  1
    Likelihood of readmission with new violent offence  0.468 ns  1
    Likelihood of new sex offence  --  1
    Completions  N = 55, X = 1655.6  N = 19, X = 1502.7
    Difference between time at risk in community  T-statistic = -0.73 ns
    Likelihood of any readmission  0.474 ns  1
    Likelihood of readmission with new offence  0.335 ns  1
    Likelihood of readmission with new violent offence  0.335 ns  1
    Likelihood of new sex offence  0.647 ns  1
    Completions  N = 1, X = 787  N = 19, X = 1502.7
    Difference between time at risk in community  T-statistic = 0.85 ns
    Likelihood of any readmission  0 ns  1
    Likelihood of readmission with new offence  0.189 ns  1
    Likelihood of readmission with new violent offence  0.189 ns  1
    Likelihood of new sex offence  0.980 ns  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Non-Aboriginal)
 Program: HI-SOP (Non-Child) (ALL)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 126  N = 73
    Change in yearly rate of violent institutional incidents (mean)  -0.056 ns  0.003 ns
    Change in yearly rate of non-violent institutional incidents (mean)  -0.571*  0.298 ns
    Likelihood of discretionary release (Odds Ratio)  2.179 ns  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 58, X = 1650.6  N = 15, X = 1551.6
    Difference between time at risk in community  T-statistic = -0.47 ns
    Likelihood of any readmission  0.212***  1
    Likelihood of readmission with new offence  0.209**  1
    Likelihood of readmission with new violent offence  0.069**  1
    Likelihood of new sex offence  --  1
    Community outcomes where approximated referral criteria met  N = 31, X = 1625.9  N = 6, X = 1396.8
    Difference between time at risk in community  T-statistic = -0.67 ns
    Likelihood of any readmission  0.564 ns  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  0 ns  1
    Completions  N = 57, X = 1664.9  N = 15, X = 1551.6
    Difference between time at risk in community  T-statistic = -0.54 ns
    Likelihood of any readmission  0.225***  1
    Likelihood of readmission with new offence  0.219*  1
    Likelihood of readmission with new violent offence  0.071**  1
    Likelihood of new sex offence  --  1
    Completions  N = 1, X = 834  N = 15, X = 1551.6
    Difference between time at risk in community  T-statistic = 0.94 ns
    Likelihood of any readmission  0 ns  1
    Likelihood of readmission with new offence  0 ns  1
    Likelihood of readmission with new violent offence  0 ns  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 Program: HI-SOP (Non-Child) (Regional)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 74  N = 21
    Change in yearly rate of violent institutional incidents (mean)  -0.065 ns  -0.005 ns
    Change in yearly rate of non-violent institutional incidents (mean)  -0.097***  -0.411 ns
    Likelihood of discretionary release (Odds Ratio)  0.605 ns  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 34, X = 1605.4  N = 3, X = 2068
    Difference between time at risk in community  T-statistic = 0.99 ns
    Likelihood of any readmission  0.276*  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Community outcomes where approximated referral criteria met  N = 23, X = 1510.7  N = 1, X = 2347
    Difference between time at risk in community  T-statistic = 1.05 ns
    Likelihood of any readmission  0.259 ns  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = 34, X = 1605.4  N = 3, X = 2068
    Difference between time at risk in community  T-statistic = 0.99 ns
    Likelihood of any readmission  0.276*  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = --, X = --  N = --, X = --
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 Program: HI-SOP (Non-Child) (Local)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 51  N = 52
    Change in yearly rate of violent institutional incidents (mean)  -0.044 ns  0.008 ns
    Change in yearly rate of non-violent institutional incidents (mean)  0.049 ns  0.585 ns
    Likelihood of discretionary release (Odds Ratio)  5.392 ns  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 24, X = 1714.5  N = 12, X = 1422.5
    Difference between time at risk in community  T-statistic = -1.21 ns
    Likelihood of any readmission  0.205***  1
    Likelihood of readmission with new offence  0.168*  1
    Likelihood of readmission with new violent offence  0.163 ns  1
    Likelihood of new sex offence  --  1
    Community outcomes where approximated referral criteria met  N = 8, X = 1957.4  N = 5, X = 1206.8
    Difference between time at risk in community  T-statistic = -2.03 *
    Likelihood of any readmission  0.364 ns  1
    Likelihood of readmission with new offence  1.000 ns  1
    Likelihood of readmission with new violent offence  1.000 ns  1
    Likelihood of new sex offence  1.000 ns  1
    Completions  N = 23, X = 1752.8  N = 12, X = 1422.5
    Difference between time at risk in community  T-statistic = -1.38 ns
    Likelihood of any readmission  0.232***  1
    Likelihood of readmission with new offence  0.186*  1
    Likelihood of readmission with new violent offence  0.184 ns  1
    Likelihood of new sex offence  --  1
    Completions  N = 1, X = 834  N = 12, X = 1422.5
    Difference between time at risk in community  T-statistic = 0.77 ns
    Likelihood of any readmission  0 ns  1
    Likelihood of readmission with new offence  0 ns  1
    Likelihood of readmission with new violent offence  0 ns  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 Program: HI-SOP (Child) (ALL)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 92  N = 40
    Change in yearly rate of violent institutional incidents (mean)  -0.006 ns  -0.213 ns
    Change in yearly rate of non-violent institutional incidents (mean)  0.063 ns  -0.011 ns
    Likelihood of discretionary release (Odds Ratio)  --  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 29, X = 1763.1  N = 3, X = 1750
    Difference between time at risk in community  T-statistic = -0.03 ns
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Community outcomes where approximated referral criteria met  N = , X =  N = , X =
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = 29, X = 1763.1  N = 3, X = 1750
    Difference between time at risk in community  T-statistic = -0.03 ns
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = --, X = --  N = --, X = --
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 Program: HI-SOP (Non-Child) (Local)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 51  N = 52
    Change in yearly rate of violent institutional incidents (mean)  -0.044 ns  0.008 ns
    Change in yearly rate of non-violent institutional incidents (mean)  0.049 ns  0.585 ns
    Likelihood of discretionary release (Odds Ratio)  5.392 ns  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 24, X = 1714.5  N = 12, X = 1422.5
    Difference between time at risk in community  T-statistic = -1.21 ns
    Likelihood of any readmission  0.205***  1
    Likelihood of readmission with new offence  0.168*  1
    Likelihood of readmission with new violent offence  0.163 ns  1
    Likelihood of new sex offence  --  1
    Community outcomes where approximated referral criteria met  N = 8, X = 1957.4  N = 5, X = 1206.8
    Difference between time at risk in community  T-statistic = -2.03 *
    Likelihood of any readmission  0.364 ns  1
    Likelihood of readmission with new offence  1.000 ns  1
    Likelihood of readmission with new violent offence  1.000 ns  1
    Likelihood of new sex offence  1.000 ns  1
    Completions  N = 23, X = 1752.8  N = 12, X = 1422.5
    Difference between time at risk in community  T-statistic = -1.38 ns
    Likelihood of any readmission  0.232***  1
    Likelihood of readmission with new offence  0.186*  1
    Likelihood of readmission with new violent offence  0.184 ns  1
    Likelihood of new sex offence  --  1
    Completions  N = 1, X = 834  N = 12, X = 1422.5
    Difference between time at risk in community  T-statistic = 0.77 ns
    Likelihood of any readmission  0 ns  1
    Likelihood of readmission with new offence  0 ns  1
    Likelihood of readmission with new violent offence  0 ns  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 Program: HI-SOP (Child) (Regional)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 59  N = 16
    Change in yearly rate of violent institutional incidents (mean)  -0.022 ns  -0.395 ns
    Change in yearly rate of non-violent institutional incidents (mean)  0.011 ns  0.996 ns
    Likelihood of discretionary release (Odds Ratio)  --  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = X =  N = X =
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Community outcomes where approximated referral criteria met  N = X =  N = X =
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = X =  N = X =
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = X =  N = X =
    Likelihood of any readmission    1
    Likelihood of readmission with new offence    1
    Likelihood of readmission with new violent offence    1
    Likelihood of new sex offence    1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 Program: HI-SOP (Child) (Local)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 33  N = 24
    Change in yearly rate of violent institutional incidents (mean)  0.021*  -0.092 ns
    Change in yearly rate of non-violent institutional incidents (mean)  0.155 ns  -0.682 ns
    Likelihood of discretionary release (Odds Ratio)  --  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 12, X = 1843.5  N = 3, X = 1750
    Difference between time at risk in community  T-statistic = -0.19 ns
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  0 ns  1
    Likelihood of readmission with new violent offence  0 ns  1
    Likelihood of new sex offence  0 ns  1
    Community outcomes where approximated referral criteria met  N = , X =  N = , X =
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = , X =  N = , X =
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  0 ns  1
    Likelihood of readmission with new violent offence  0 ns  1
    Likelihood of new sex offence  0 ns  1
    Completions  N = , X =  N = , X =
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (non-Aboriginal)
 Program: MI-SOP (Non-Child) (All)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 500  N = 295
    Change in yearly rate of violent institutional incidents (mean)  -0.012 ns  -0.016 ns
    Change in yearly rate of non-violent institutional incidents (mean)  0.116 ns  0.002 ns
    Likelihood of discretionary release (Odds Ratio)  5.763***  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 327, X = 1733.6  N = 128, X = 1831
    Difference between time at risk in community  T-statistic = 1.16 ns
    Likelihood of any readmission  0.721*  1
    Likelihood of readmission with new offence  0.559 ns  1
    Likelihood of readmission with new violent offence  0.457 ns  1
    Likelihood of new sex offence  0.298 ns  1
    Community outcomes where approximated referral criteria met  N = 93, X = 1759.2  N = 41, X = 1709
    Difference between time at risk in community  T-statistic = -0.33 ns
    Likelihood of any readmission  0.710 ns  1
    Likelihood of readmission with new offence  0.365 ns  1
    Likelihood of readmission with new violent offence  0.201 ns  1
    Likelihood of new sex offence  0.201 ns  1
    Completions  N = 316, X = 1729.6  N = 128, X = 1831
    Difference between time at risk in community  T-statistic = 1.21 ns
    Likelihood of any readmission  0.713*  1
    Likelihood of readmission with new offence  0.568 ns  1
    Likelihood of readmission with new violent offence  0.465 ns  1
    Likelihood of new sex offence  0.303 ns  1
    Completions  N = 11, X = 1848.6  N = 128, X = 1831
    Difference between time at risk in community  T-statistic = -0.07 ns
    Likelihood of any readmission  0.822 ns  1
    Likelihood of readmission with new offence  0 ns  1
    Likelihood of readmission with new violent offence  0 ns  1
    Likelihood of new sex offence  0 ns  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (non-Aboriginal)
 Program: MI-SOP (non-Child) (National)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 232  N = 163
    Change in yearly rate of violent institutional incidents (mean)  -0.002 ns  -0.026 ns
    Change in yearly rate of non-violent institutional incidents (mean)  0.187 ns  0.341 ns
    Likelihood of discretionary release (Odds Ratio)  5.072***  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 139, X = 1570.1  N = 54, X = 1570.7
    Difference between time at risk in community  T-statistic = 0 ns
    Likelihood of any readmission  0.549**  1
    Likelihood of readmission with new offence  0.444 ns  1
    Likelihood of readmission with new violent offence  0.134 ns  1
    Likelihood of new sex offence  0 ns  1
    Community outcomes where approximated referral criteria met  N = 46, X = 1628.5  N = 23, X = 1476.3
    Difference between time at risk in community  T-statistic = -0.71 ns
    Likelihood of any readmission  0.637 ns  1
    Likelihood of readmission with new offence  0.365 ns  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = 132, X = 1550.7  N = 54, X = 1570.7
    Difference between time at risk in community  T-statistic = 0.16 ns
    Likelihood of any readmission  0.538**  1
    Likelihood of readmission with new offence  0.456 ns  1
    Likelihood of readmission with new violent offence  0.137 ns  1
    Likelihood of new sex offence  0 ns  1
    Completions  N = 7, X = 1936.3  N = 54, X = 1570.7
    Difference between time at risk in community  T-statistic = -1.10 ns
    Likelihood of any readmission  0.753 ns  1
    Likelihood of readmission with new offence  0 ns  1
    Likelihood of readmission with new violent offence  0 ns  1
    Likelihood of new sex offence  0 ns  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (non-Aboriginal)
 Program: MI-SOP (non-Child) (Regional)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 21  N = 25
    Change in yearly rate of violent institutional incidents (mean)  0.122 ns  -0.027 ns
    Change in yearly rate of non-violent institutional incidents (mean)  0.430 ns  -0.78 ns
    Likelihood of discretionary release (Odds Ratio)  1.675 ns  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 10, X = 2170.6  N = 10, X = 2263
    Difference between time at risk in community  T-statistic = 0.38 ns
    Likelihood of any readmission  1.653 ns  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Community outcomes where approximated referral criteria met  N = 5, X = 2200.4  N = 4, X = 2265
    Difference between time at risk in community  T-statistic = 0.20 ns
    Likelihood of any readmission  1.693 ns  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = 10, X = 2170.6  N = 10, X = 2263
    Difference between time at risk in community  T-statistic = 0.38 ns
    Likelihood of any readmission  1.653 ns  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = --, X = --  N = --, X = --
    Difference between time at risk in community  T-statistic = --
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (non-Aboriginal)
 Program: MI-SOP (non-Child) (Local)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 247  N = 107
    Change in yearly rate of violent institutional incidents (mean)  -0.033** ns  0.003 ns
    Change in yearly rate of non-violent institutional incidents (mean)  0.021 ns  -0.331 ns
    Likelihood of discretionary release (Odds Ratio)  6.529***  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 178, X = 1836.6  N = 64, X = 1983.2
    Difference between time at risk in community  T-statistic = 1.22 ns
    Likelihood of any readmission  0.831 ns  1
    Likelihood of readmission with new offence  0.610 ns  1
    Likelihood of readmission with new violent offence  0.722 ns  1
    Likelihood of new sex offence  0.441 ns  1
    Community outcomes where approximated referral criteria met  N = 42, X = 1849.9  N = 14, X = 1932.4
    Difference between time at risk in community  T-statistic = 0.35 ns
    Likelihood of any readmission  0.895 ns  1
    Likelihood of readmission with new offence  0.333 ns  1
    Likelihood of readmission with new violent offence  0.148 ns  1
    Likelihood of new sex offence  0.148 ns  1
    Completions  N = 174, X = 1839.9  N = 64, X = 1983.2
    Difference between time at risk in community  T-statistic = 1.19 ns
    Likelihood of any readmission  0.834 ns  1
    Likelihood of readmission with new offence  0.615 ns  1
    Likelihood of readmission with new violent offence  0.726 ns  1
    Likelihood of new sex offence  0.444 ns  1
    Completions  N = 4, X = 1695.3  N = 64, X = 1983.2
    Difference between time at risk in community  T-statistic = 0.72 ns
    Likelihood of any readmission  0.416 ns  1
    Likelihood of readmission with new offence  0 ns  1
    Likelihood of readmission with new violent offence  0 ns  1
    Likelihood of new sex offence  0 ns  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (non-Aboriginal)
 Program: MI-SOP (Child) (All)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 706  N = 227
    Change in yearly rate of violent institutional incidents (mean)  0.002 ns  -0.034*
    Change in yearly rate of non-violent institutional incidents (mean)  0.045 ns  0.056 ns
    Likelihood of discretionary release (Odds Ratio)  11.222***  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 497, X = 1777.3  N = 76, X = 1815.9
    Difference between time at risk in community  T-statistic = 0.39 ns
    Likelihood of any readmission  0.731 ns  1
    Likelihood of readmission with new offence  0.640 ns  1
    Likelihood of readmission with new violent offence  1.486 ns  1
    Likelihood of new sex offence  0.747 ns  1
    Community outcomes where approximated referral criteria met  N = 113, X = 1644  N = 31, X = 1707.3
    Difference between time at risk in community  T-statistic = 0.42 ns
    Likelihood of any readmission  0.990 ns  1
    Likelihood of readmission with new offence  0.854 ns  1
    Likelihood of readmission with new violent offence  1.024 ns  1
    Likelihood of new sex offence  0.650 ns  1
    Completions  N = 481, X = 1790.2  N = 76, X = 1815.9
    Difference between time at risk in community  T-statistic = 0.26 ns
    Likelihood of any readmission  0.747 ns  1
    Likelihood of readmission with new offence  0.658 ns  1
    Likelihood of readmission with new violent offence  1.543 ns  1
    Likelihood of new sex offence  0.785 ns  1
    Completions  N = 16, X = 1387.2  N = 76, X = 1815.9
    Difference between time at risk in community  T-statistic = 2.20**
    Likelihood of any readmission  0.339 ns  1
    Likelihood of readmission with new offence  0 ns  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Remarque : rapport de risques; ns = non significatif; *p < 0,10, **p < 0,05, ***p < 0,01.


 Effectiveness Indicator (non-Aboriginal)
 Program: MI-SOP (Child) (National)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 315  N = 129
    Change in yearly rate of violent institutional incidents (mean)  0.009 ns  -0.016 ns
    Change in yearly rate of non-violent institutional incidents (mean)  -0.008 ns  0.195 ns
    Likelihood of discretionary release (Odds Ratio)  10.695***  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 225, X = 1716.3  N = 33, X = 1733.8
    Difference between time at risk in community  T-statistic = 0.12 ns
    Likelihood of any readmission  1.037 ns  1
    Likelihood of readmission with new offence  1.665 ns  1
    Likelihood of readmission with new violent offence  0.664 ns  1
    Likelihood of new sex offence  0.192 ns  1
    Community outcomes where approximated referral criteria met  N = 56, X = 1529  N = 17, X = 1534.1
    Difference between time at risk in community  T-statistic = 0.03 ns
    Likelihood of any readmission  1.975 ns  1
    Likelihood of readmission with new offence  3.379 ns  1
    Likelihood of readmission with new violent offence  0.703 ns  1
    Likelihood of new sex offence  0.505 ns  1
    Completions  N = 217, X = 1726.7  N = 33, X = 1733.8
    Difference between time at risk in community  T-statistic = 0.05 ns
    Likelihood of any readmission  1.094 ns  1
    Likelihood of readmission with new offence  1.745 ns  1
    Likelihood of readmission with new violent offence  0.703 ns  1
    Likelihood of new sex offence  0.197 ns  1
    Completions  N = 8, X = 1434.4  N = 33, X = 1733.8
    Difference between time at risk in community  T-statistic = 1.01 ns
    Likelihood of any readmission  0 ns  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (non-Aboriginal)
 Program: MI-SOP (Child) (Regional)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 21  N = 12
    Change in yearly rate of violent institutional incidents (mean)  0.010 ns  -0.081 ns
    Change in yearly rate of non-violent institutional incidents (mean)  0.651 ns  -0.168 ns
    Likelihood of discretionary release (Odds Ratio)  --  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 8, X = 2087.8  N = 4, X = 2490.5
    Difference between time at risk in community  T-statistic = 1.16 ns
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Community outcomes where approximated referral criteria met  N = 4, X = 2255  N = 2, X = 2277.5
    Difference between time at risk in community  T-statistic = 0.05 ns
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = 7, X = 2172.9  N = 4, X = 2490.5
    Difference between time at risk in community  T-statistic = 0.91 ns
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = 1, X = 1492  N = 4, X = 2490.5
    Difference between time at risk in community  T-statistic = 2.02 ns
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (non-Aboriginal)
 Program: MI-SOP (Child) (Local)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 370  N = 86
    Change in yearly rate of violent institutional incidents (mean)  -0.003 ns  -0.054**
    Change in yearly rate of non-violent institutional incidents (mean)  0.055 ns  -0.122 ns
    Likelihood of discretionary release (Odds Ratio)  20.468***  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 264, X = 1819.8  N = 39, X = 1816.1
    Difference between time at risk in community  T-statistic = -0.03 ns
    Likelihood of any readmission  0.404**  1
    Likelihood of readmission with new offence  0.199**  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Community outcomes where approximated referral criteria met  N = 53, X = 1719.4  N = 12, X = 1857.5
    Difference between time at risk in community  T-statistic = 0.57 ns
    Likelihood of any readmission  0.261**  1
    Likelihood of readmission with new offence  0.113**  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = 257, X = 1833.5  N = 39, X = 1816.1
    Difference between time at risk in community  T-statistic = -0.12 ns
    Likelihood of any readmission  0.404**  1
    Likelihood of readmission with new offence  0.202**  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = 7, X = 1318.7  N = 39, X = 1816.1
    Difference between time at risk in community  T-statistic = 1.79*
    Likelihood of any readmission  0.320 ns  1
    Likelihood of readmission with new offence  0 ns  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 Program: MI-SOP (Non-Child) (All)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 172  N = 104
    Change in yearly rate of violent institutional incidents (mean)  -0.021*  0.066 ns
    Change in yearly rate of non-violent institutional incidents (mean)  0.592**  -0.446 ns
    Likelihood of discretionary release (Odds Ratio)  1.862 ns  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 101, X = 1764.2  N = 36, X = 1621.6
    Difference between time at risk in community  T-statistic = -0.91 ns
    Likelihood of any readmission  0.793 ns  1
    Likelihood of readmission with new offence  0.667 ns  1
    Likelihood of readmission with new violent offence  2.496 ns  1
    Likelihood of new sex offence  0.897 ns  1
    Community outcomes where approximated referral criteria met  N = 53, X = 1781.7  N = 15, X = 1423.6
    Difference between time at risk in community  T-statistic = -1.47 ns
    Likelihood of any readmission  1.112 ns  1
    Likelihood of readmission with new offence  1.307 ns  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  1.387 ns  1
    Completions  N = 94, X = 1746.4  N = 36, X = 1621.6
    Difference between time at risk in community  T-statistic = -0.78 ns
    Likelihood of any readmission  0.754 ns  1
    Likelihood of readmission with new offence  0.613 ns  1
    Likelihood of readmission with new violent offence  2.736 ns  1
    Likelihood of new sex offence  0.968 ns  1
    Completions  N = 7, X = 2002.3  N = 36, X = 1621.6
    Difference between time at risk in community  T-statistic = -1.21 ns
    Likelihood of any readmission  1.446 ns  1
    Likelihood of readmission with new offence  1.206 ns  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  9.583 ns  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 Program: MI-SOP (Non-Child) (National)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 65  N = 47
    Change in yearly rate of violent institutional incidents (mean)  -0.01 ns  -0.012 ns
    Change in yearly rate of non-violent institutional incidents (mean)  0.757 ns  -0.625*
    Likelihood of discretionary release (Odds Ratio)  1.134 ns  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 37, X = 1591.8  N = 13, X = 1467.2
    Difference between time at risk in community  T-statistic = -0.52 ns
    Likelihood of any readmission  1.019 ns  1
    Likelihood of readmission with new offence  0.552 ns  1
    Likelihood of readmission with new violent offence  0 ns  1
    Likelihood of new sex offence  0 ns  1
    Community outcomes where approximated referral criteria met  N = 19, X = 1605.8  N = 5, X = 1589.2
    Difference between time at risk in community  T-statistic = -0.04 ns
    Likelihood of any readmission  4.172 ns  1
    Likelihood of readmission with new offence  1.869 ns  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = 33, X = 1554.5  N = 13, X = 1467.2
    Difference between time at risk in community  T-statistic = -0.36 ns
    Likelihood of any readmission  0.770 ns  1
    Likelihood of readmission with new offence  0.379 ns  1
    Likelihood of readmission with new violent offence  0 ns  1
    Likelihood of new sex offence  0 ns  1
    Completions  N = 4, X = 1899.8  N = 13, X = 1467.2
    Difference between time at risk in community  T-statistic = -0.91 ns
    Likelihood of any readmission  2.757 ns  1
    Likelihood of readmission with new offence  2.490 ns  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 Program: MI-SOP (Non-Child) (Regional)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 3  N = 6
    Change in yearly rate of violent institutional incidents (mean)  -0.388 ns  0.385 ns
    Change in yearly rate of non-violent institutional incidents (mean)  1.362 ns  -0.19 ns
    Likelihood of discretionary release (Odds Ratio)  --  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 1, X = 2520  N = 2, X = 1412
    Difference between time at risk in community  T-statistic = -1.15 ns
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Community outcomes where approximated referral criteria met  N = 1, X = 2520  N = 1, X = 1967
    Difference between time at risk in community  T-statistic = -- ns
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = 1, X = 2520  N = 2, X = 1412
    Difference between time at risk in community  T-statistic = -1.15 ns
    Likelihood of any readmission  0 ns  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = --, X = --  N = --, X = --
    Difference between time at risk in community  T-statistic = --
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 Program: MI-SOP (Non-Child) (Local)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 104  N = 51
    Change in yearly rate of violent institutional incidents (mean)  -0.018 ns  0.102 ns
    Change in yearly rate of non-violent institutional incidents (mean)  0.467*  -0.311 ns
    Likelihood of discretionary release (Odds Ratio)  2.062 ns  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 63, X = 1853.4  N = 21, X = 1737.1
    Difference between time at risk in community  T-statistic = -0.55 ns
    Likelihood of any readmission  0.714 ns  1
    Likelihood of readmission with new offence  0.749 ns  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Community outcomes where approximated referral criteria met  N = 33, X = 1860.6  N = 9, X = 1271.2
    Difference between time at risk in community  T-statistic = -1.75*
    Likelihood of any readmission  0.775 ns  1
    Likelihood of readmission with new offence  1.258 ns  1
    Likelihood of readmission with new violent offence  2.89 ns  1
    Likelihood of new sex offence  2.488 ns  1
    Completions  N = 60, X = 1839.1  N = 21, X = 1737.1
    Difference between time at risk in community  T-statistic = -0.47 ns
    Likelihood of any readmission  0.759 ns  1
    Likelihood of readmission with new offence  0.781 ns  1
    Likelihood of readmission with new violent offence  4.147 ns  1
    Likelihood of new sex offence  6.375 ns  1
    Completions  N = 3, X = 2139  N = 21, X = 1737.1
    Difference between time at risk in community  T-statistic = -0.87 ns
    Likelihood of any readmission  0.357 ns  1
    Likelihood of readmission with new offence  0.569 ns  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 Program: MI-SOP (Child) (All)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 123  N = 52
    Change in yearly rate of violent institutional incidents (mean)  -0.057*  0.034 ns
    Change in yearly rate of non-violent institutional incidents (mean)  0.072 ns  0.083 ns
    Likelihood of discretionary release (Odds Ratio)  3.907 ns  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 77, X = 1858.3  N = 19, X = 1495.4
    Difference between time at risk in community  T-statistic = -1.61 ns
    Likelihood of any readmission  1.292 ns  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Community outcomes where approximated referral criteria met  N = 29, X = 1867.5  N = 5, X = 1626.8
    Difference between time at risk in community  T-statistic = -0.56 ns
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = 74, X = 1846.3  N = 19, X = 1495.4
    Difference between time at risk in community  T-statistic = -1.54 ns
    Likelihood of any readmission  1.260 ns  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  1,668 ns  1
    Completions  N = 3, X = 2154.7  N = 19, X = 1495.4
    Difference between time at risk in community  T-statistic = 9.108 ns
    Likelihood of any readmission  5.252 ns  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 Program: MI-SOP (Child) (National)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 40  N = 22
    Change in yearly rate of violent institutional incidents (mean)  -0.094 ns  -0.043 ns
    Change in yearly rate of non-violent institutional incidents (mean)  -0.867 ns  -0.433 ns
    Likelihood of discretionary release (Odds Ratio)  23.510 ns  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 24, X = 1691.1  N = 6, X = 855.67
    Difference between time at risk in community  T-statistic = -2.20**
    Likelihood of any readmission  0.338 ns  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  0 ns  1
    Community outcomes where approximated referral criteria met  N = 8, X = 1910.3  N = 1, X = 388
    Difference between time at risk in community  T-statistic = -1.60 ns
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = 23, X = 1685.5  N = 6, X = 855.67
    Difference between time at risk in community  T-statistic = -2.14**
    Likelihood of any readmission  0.356 ns  1
    Likelihood of readmission with new offence  4.571 ns  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = 1, X = 1820  N = 6, X = 855.67
    Difference between time at risk in community  T-statistic = -1.85 ns
    Likelihood of any readmission  0 ns  1
    Likelihood of readmission with new offence  2.398 ns  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 Program: MI-SOP (Child) (Regional)
 Intent to Treat  Comparison
 Institutional Outcomes  N = --  N = 2
    Change in yearly rate of violent institutional incidents (mean)  --  --
    Change in yearly rate of non-violent institutional incidents (mean)  --  2.855 ns
    Likelihood of discretionary release (Odds Ratio)    1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = --, X = --  N = --, X = --
    Difference between time at risk in community  T-statistic = --
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Community outcomes where approximated referral criteria met  N = --, X = --  N = --, X = --
    Difference between time at risk in community  T-statistic = --
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = --, X = --  N = --, X = --
    Difference between time at risk in community  T-statistic = --
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = --, X = --  N = --, X = --
    Difference between time at risk in community T-statistic = --
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 Program: MI-SOP (Child) (Local)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 82  N = 28
    Change in yearly rate of violent institutional incidents (mean)  -0.04 ns  0.096 ns
    Change in yearly rate of non-violent institutional incidents (mean)  0.532**  0.291 ns
    Likelihood of discretionary release (Odds Ratio)  2.658 ns  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 53, X = 1934  N = 12, X = 1804
    Difference between time at risk in community  T-statistic = -0.46 ns
    Likelihood of any readmission  2.530 ns  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Community outcomes where approximated referral criteria met  N = 21, X = 1851.2  N = 4, X = 1936.5
    Difference between time at risk in community  T-statistic = 0.18 ns
    Likelihood of any readmission  1.970 ns  1
    Likelihood of readmission with new offence  4.138 ns  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = 51, X = 1918.8  N = 12, X = 1804
    Difference between time at risk in community  T-statistic = -0.40 ns
    Likelihood of any readmission  2.494 ns  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  2.192 ns  1
    Completions  N = 2, X = 2322  N = 12, X = 1804
    Difference between time at risk in community  T-statistic = -0.94 ns
    Likelihood of any readmission  5.926 ns  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (non-Aboriginal)
 Program: LI-SOP (Non-Child) (All)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 158  N = 57
    Change in yearly rate of violent institutional incidents (mean)  -0.031 ns  0.210 ns
    Change in yearly rate of non-violent institutional incidents (mean)  -0.143 ns  -0.193 ns
    Likelihood of discretionary release (Odds Ratio)  3.467***  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 141, X = 2043.7  N = 35, X = 1666.4
    Difference between time at risk in community  T-statistic = -2.37**
    Likelihood of any readmission  0.592 ns  1
    Likelihood of readmission with new offence  0.136***  1
    Likelihood of readmission with new violent offence  0.086*  1
    Likelihood of new sex offence  0.098 ns  1
    Community outcomes where approximated referral criteria met  N = 17, X = 2334.5  N = 11, X = 1543.6
    Difference between time at risk in community  T-statistic = -2.54**
    Likelihood of any readmission  0.181**  1
    Likelihood of readmission with new offence  0 ns  1
    Likelihood of readmission with new violent offence  0 ns  1
    Likelihood of new sex offence  --  1
    Completions  N = 139, X = 2044.8  N = 35, X = 1666.4
    Difference between time at risk in community  T-statistic = -2.36**
    Likelihood of any readmission  0.598 ns  1
    Likelihood of readmission with new offence  0.139***  1
    Likelihood of readmission with new violent offence  0.086*  1
    Likelihood of new sex offence  0.098 ns  1
    Completions  N = 2, X = 1973.5  N = 35, X = 1666.4
    Difference between time at risk in community  T-statistic = -0.55 ns
    Likelihood of any readmission  0 ns  1
    Likelihood of readmission with new offence  0 ns  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (non-Aboriginal)
 Program: LI-SOP (non-Child) (National)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 79  N = 35
    Change in yearly rate of violent institutional incidents (mean)  --  0.349 ns
    Change in yearly rate of non-violent institutional incidents (mean)  -0.007 ns  0.051 ns
    Likelihood of discretionary release (Odds Ratio)  4.108*  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 59, X = 1327.8  N = 14, X = 1059.7
    Difference between time at risk in community  T-statistic = -1.60 ns
    Likelihood of any readmission  0.409 ns  1
    Likelihood of readmission with new offence  0.237 ns  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Community outcomes where approximated referral criteria met  N = 7, X = 1525.4  N = 7, X = 1229.7
    Difference between time at risk in community  T-statistic = -1.07 ns
    Likelihood of any readmission  0.020**  1
    Likelihood of readmission with new offence  0 ns  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = 59, X = 1327.8  N = 14, X = 1059.7
    Difference between time at risk in community  T-statistic = -1.60 ns
    Likelihood of any readmission  0.409 ns  1
    Likelihood of readmission with new offence  0.237 ns  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = --, X = --  N = --, X = --
    Difference between time at risk in community  T-statistic = --
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (non-Aboriginal)
 Program: LI-SOP (non-Child) (Regional)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 33  N = 2
    Change in yearly rate of violent institutional incidents (mean)  --  --
    Change in yearly rate of non-violent institutional incidents (mean)  0.207 ns  --
    Likelihood of discretionary release (Odds Ratio)  --  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 32, X = 2247.5  N = 2, X = 1458
    Difference between time at risk in community  T-statistic = -1.54 ns
    Likelihood of any readmission  0.162 ns  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  2.025 ns  1
    Likelihood of new sex offence  2.025 ns  1
    Community outcomes where approximated referral criteria met  N = --, X = --  N = --, X = --
    Difference between time at risk in community  T-statistic = --
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = 31, X = 2245.4  N = 2, X = 1458
    Difference between time at risk in community  T-statistic = -1.51 ns
    Likelihood of any readmission  0.178 ns  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  2.353 ns  1
    Likelihood of new sex offence  2.353 ns  1
    Completions  N = 1, X = 2313  N = 2, X = 1458
    Difference between time at risk in community  T-statistic = -2.54 ns
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (non-Aboriginal)
 Program: LI-SOP (non-Child) (Local)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 46  N = 20
    Change in yearly rate of violent institutional incidents (mean)  -0.107 ns  -0.012 ns
    Change in yearly rate of non-violent institutional incidents (mean)  -0.627 ns  -0.639*
    Likelihood of discretionary release (Odds Ratio)  2.835*  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 50, X = 2758.2  N = 19, X = 2135.3
    Difference between time at risk in community  T-statistic = -4.31***
    Likelihood of any readmission  0.838 ns  1
    Likelihood of readmission with new offence  0.101**  1
    Likelihood of readmission with new violent offence  0 ns  1
    Likelihood of new sex offence  0 ns  1
    Community outcomes where approximated referral criteria met  N = 9, X = 2889.8  N = 4, X = 2093
    Difference between time at risk in community  T-statistic = -2.24**
    Likelihood of any readmission  0.159*  1
    Likelihood of readmission with new offence  0 ns  1
    Likelihood of readmission with new violent offence  0 ns  1
    Likelihood of new sex offence  --  1
    Completions  N = 49, X = 2781.1  N = 19, X = 2135.3
    Difference between time at risk in community  T-statistic = -4.58***
    Likelihood of any readmission  0.836 ns  1
    Likelihood of readmission with new offence  0.103**  1
    Likelihood of readmission with new violent offence  0 ns  1
    Likelihood of new sex offence  0 ns  1
    Completions  N = 1, X = 1634  N = 19, X = 2135.3
    Difference between time at risk in community  T-statistic = 0.70 ns
    Likelihood of any readmission  0.276 ns  1
    Likelihood of readmission with new offence  0.094 ns  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (non-Aboriginal)
 Program: LI-SOP (Child) (All)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 403  N = 57
    Change in yearly rate of violent institutional incidents (mean)  --  0.082 ns
    Change in yearly rate of non-violent institutional incidents (mean)  0.051 ns  0.104 ns
    Likelihood of discretionary release (Odds Ratio)  8.142***  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 339, X = 1917.6  N = 45, X = 1660.6
    Difference between time at risk in community  T-statistic = -1.86*
    Likelihood of any readmission  0.290***  1
    Likelihood of readmission with new offence  0.309 ns  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Community outcomes where approximated referral criteria met  N = 29, X = 2064.4  N = 7, X = 1825.3
    Difference between time at risk in community  T-statistic = -0.67 ns
    Likelihood of any readmission  0.052***  1
    Likelihood of readmission with new offence  0 ns  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = 335, X = 1920  N = 45, X = 1660.6
    Difference between time at risk in community  T-statistic = -1.87*
    Likelihood of any readmission  0.290***  1
    Likelihood of readmission with new offence  0.310 ns  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = 4, X = 1717  N = 45, X = 1660.6
    Difference between time at risk in community  T-statistic = -0.12 ns
    Likelihood of any readmission  0 nss  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (non-Aboriginal)
 Program: LI-SOP (Child) (National)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 215  N = 29
    Change in yearly rate of violent institutional incidents (mean)  --  --
    Change in yearly rate of non-violent institutional incidents (mean)  0.063 ns  -0.099 ns
    Likelihood of discretionary release (Odds Ratio)  17.764***  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 147, X = 1237.6  N = 20, X = 878.7
    Difference between time at risk in community  T-statistic = -2.64***
    Likelihood of any readmission  0.219***  1
    Likelihood of readmission with new offence  0.334 ns  1
    Likelihood of readmission with new violent offence  7.115 ns  1
    Likelihood of new sex offence  --  1
    Community outcomes where approximated referral criteria met  N = 8, X = 1468.8  N = 4, X = 1126
    Difference between time at risk in community  T-statistic = -0.82 ns
    Likelihood of any readmission  0 ns  1
    Likelihood of readmission with new offence  0 ns  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = 147, X = 1237.6  N = 20, X = 878.7
    Difference between time at risk in community  T-statistic = -2.64***
    Likelihood of any readmission  0.219***  1
    Likelihood of readmission with new offence  0.334 ns  1
    Likelihood of readmission with new violent offence  7.115 ns  1
    Likelihood of new sex offence  --  1
    Completions  N = --, X = --  N = --, X = --
    Difference between time at risk in community  T-statistic = --
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (non-Aboriginal)
 Program: LI-SOP (Child) (Regional)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 128  N = 4
    Change in yearly rate of violent institutional incidents (mean)  --  --
    Change in yearly rate of non-violent institutional incidents (mean)  -0.008 ns  --
    Likelihood of discretionary release (Odds Ratio)  19.449 ns  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 133, X = 2252.5  N = 5, X = 1681
    Difference between time at risk in community  T-statistic = -1.88*
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Community outcomes where approximated referral criteria met  N = --, X = --  N = --, X = --
    Difference between time at risk in community  T-statistic = --
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = 130, X = 2268.6  N = 5, X = 1681
    Difference between time at risk in community  T-statistic = -1.94*
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = 3, X = 1553.7  N = 5, X = 1681
    Difference between time at risk in community  T-statistic = 0.23 ns
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (non-Aboriginal)
 Program: LI-SOP (Child) (Local)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 60  N = 24
    Change in yearly rate of violent institutional incidents (mean)  --  0.195 ns
    Change in yearly rate of non-violent institutional incidents (mean)  0.130 ns  0.367 ns
    Likelihood of discretionary release (Odds Ratio)  15.403***  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 59, X = 2857.1  N = 20, X = 2437.5
    Difference between time at risk in community  T-statistic = -3.54***
    Likelihood of any readmission  0.317**  1
    Likelihood of readmission with new offence  0.475 ns  1
    Likelihood of readmission with new violent offence  1.255 ns  1
    Likelihood of new sex offence  1.255 ns  1
    Community outcomes where approximated referral criteria met  N = 7, X = 2854.7  N = 3, X = 2757.7
    Difference between time at risk in community  T-statistic = -0.35 ns
    Likelihood of any readmission  6.045 ns  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = 58, X = 2868.3  N = 20, X = 2437.5
    Difference between time at risk in community  T-statistic = -3.66***
    Likelihood of any readmission  0.317**  1
    Likelihood of readmission with new offence  0.475 ns  1
    Likelihood of readmission with new violent offence  1.255 ns  1
    Likelihood of new sex offence  1.255 ns  1
    Completions  N = 1, X = 2207  N = 20, X = 2437.5
    Difference between time at risk in community  T-statistic = 0.40 ns
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 rogram: LI-SOP (Non-Child) (All)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 58  N = 37
    Change in yearly rate of violent institutional incidents (mean)  0.022 ns  0.036 ns
    Change in yearly rate of non-violent institutional incidents (mean)  0.213 ns  -0.663 ns
    Likelihood of discretionary release (Odds Ratio)  24.881***  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 58, X = 1841.3  N = 20, X = 1573.8
    Difference between time at risk in community  T-statistic = -1.18 ns
    Likelihood of any readmission  0.449*  1
    Likelihood of readmission with new offence  0.430 ns  1
    Likelihood of readmission with new violent offence  1.448 ns  1
    Likelihood of new sex offence  0 ns  1
    Community outcomes where approximated referral criteria met  N = 18, X = 2262.3  N = 13, X = 1531.5
    Difference between time at risk in community  T-statistic = -2.37**
    Likelihood of any readmission  0.415*  1
    Likelihood of readmission with new offence  0.309 ns  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  0 ns  1
    Completions  N = 58, X = 1841.3  N = 20, X = 1573.8
    Difference between time at risk in community  T-statistic = -1.18 ns
    Likelihood of any readmission  0.449*  1
    Likelihood of readmission with new offence  0.430 ns  1
    Likelihood of readmission with new violent offence  1.448 ns  1
    Likelihood of new sex offence  0 ns  1
    Completions  N = --, X = --  N = --, X = --
    Difference between time at risk in community  T-statistic = --
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 Program: LI-SOP (Non-Child) (National)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 37  N = 28
    Change in yearly rate of violent institutional incidents (mean)  -0.019 ns  -0.005 ns
    Change in yearly rate of non-violent institutional incidents (mean)  0.148 ns  -1.393*
    Likelihood of discretionary release (Odds Ratio)  68.981***  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 31, X = 1355.6  N = 15, X = 1241.2
    Difference between time at risk in community  T-statistic = -0.63 ns
    Likelihood of any readmission  0,358*  1
    Likelihood of readmission with new offence  0.324 ns  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Community outcomes where approximated referral criteria met  N = 5, X = 1322.6  N = 10, X = 1262.2
    Difference between time at risk in community  T-statistic = -0.18 ns
    Likelihood of any readmission  0.245 ns  1
    Likelihood of readmission with new offence  0.314 ns  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = 31, X = 1355.6  N = 15, X = 1241.2
    Difference between time at risk in community  T-statistic = -0.63 ns
    Likelihood of any readmission  0.358*  1
    Likelihood of readmission with new offence  0.324 ns  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = --, X = --  N = --, X = --
    Difference between time at risk in community  T-statistic = --
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 Program: LI-SOP (Non-Child) (Regional)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 4  N = --
    Change in yearly rate of violent institutional incidents (mean)  -0.042 ns  --
    Change in yearly rate of non-violent institutional incidents (mean)  1.005 ns  --
    Likelihood of discretionary release (Odds Ratio)  --  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = --, X = --  N = --, X = --
    Difference between time at risk in community  T-statistic = --
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Community outcomes where approximated referral criteria met  N = --, X = --  N = --, X = --
    Difference between time at risk in community  T-statistic = --
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = --, X = --  N = --, X = --
    Difference between time at risk in community  T-statistic = --
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = --, X = --  N = --, X = --
    Difference between time at risk in community  T-statistic = --
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 Program: LI-SOP (Non-Child) (Local)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 17  N = 9
    Change in yearly rate of violent institutional incidents (mean)  0.125 ns  0.163 ns
    Change in yearly rate of non-violent institutional incidents (mean)  0.168 ns  1.606 ns
    Likelihood of discretionary release (Odds Ratio)  --  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 25, X = 2492.7  N = 5, X = 2571.6
    Difference between time at risk in community  T-statistic = 0.20 ns
    Likelihood of any readmission  0.921 ns  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Community outcomes where approximated referral criteria met  N = 11, X = 2877.5  N = 3, X = 2429.3
    Difference between time at risk in community  T-statistic = -1.69 ns
    Likelihood of any readmission  2.014 ns  1
    Likelihood of readmission with new offence  1.061 ns  1
    Likelihood of readmission with new violent offence  3.654 ns  1
    Likelihood of new sex offence  0 ns  1
    Completions  N = 25, X = 2492.7  N = 5, X = 2751.6
    Difference between time at risk in community  T-statistic = 0.20 ns
    Likelihood of any readmission  0.921 ns  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  4.938 ns  1
    Likelihood of new sex offence  0 ns  1
    Completions  N = --, X = --  N = --, X = --
    Difference between time at risk in community  T-statistic = --
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 Program: LI-SOP (Child) (All)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 56  N = 17
    Change in yearly rate of violent institutional incidents (mean)  -0.011 ns  --
    Change in yearly rate of non-violent institutional incidents (mean)  0.102 ns  -0.338 ns
    Likelihood of discretionary release (Odds Ratio)  62.008**  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 50, X = 1710  N = 11, X = 1740.7
    Difference between time at risk in community  T-statistic = 0.11 ns
    Likelihood of any readmission  0.350**  1
    Likelihood of readmission with new offence  0.285 ns  1
    Likelihood of readmission with new violent offence  0.089*  1
    Likelihood of new sex offence  0.089*  1
    Community outcomes where approximated referral criteria met  N = 10, X = 2166  N = 2, X = 2250.5
    Difference between time at risk in community  T-statistic = 0.13 ns
    Likelihood of any readmission  0 ns  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = 50, X = 1710  N = 11, X = 1740.7
    Difference between time at risk in community  T-statistic = 0.11 ns
    Likelihood of any readmission  0.350**  1
    Likelihood of readmission with new offence  0.285 ns  1
    Likelihood of readmission with new violent offence  0,089*  1
    Likelihood of new sex offence  0.089*  1
    Completions  N = --, X = --  N = --, X = --
    Difference between time at risk in community  T-statistic = --
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 Program: LI-SOP (Child) (National)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 37  N = 11
    Change in yearly rate of violent institutional incidents (mean)  --  --
    Change in yearly rate of non-violent institutional incidents (mean)  0.155 ns  0.108 ns
    Likelihood of discretionary release (Odds Ratio)  --  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 29, X = 1382.2  N = 5, X = 1360.2
    Difference between time at risk in community  T-statistic = -0.07 ns
    Likelihood of any readmission  0.112***  1
    Likelihood of readmission with new offence  0.045**  1
    Likelihood of readmission with new violent offence  0.004 ns  1
    Likelihood of new sex offence  0.004 ns  1
    Community outcomes where approximated referral criteria met  N = 4, X = 1899.5  N = 1, X = 1876
    Difference between time at risk in community  T-statistic = -0.04 ns
    Likelihood of any readmission  0 ns  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = 29, X = 1382.2  N = 5, X = 1360.2
    Difference between time at risk in community  T-statistic = -0.07 ns
    Likelihood of any readmission  0.112***  1
    Likelihood of readmission with new offence  0.045**  1
    Likelihood of readmission with new violent offence  0.004 ns  1
    Likelihood of new sex offence  0.004 ns  1
    Completions  N = --, X = --  N = --, X = --
    Difference between time at risk in community  T-statistic = --
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 Program: LI-SOP (Child) (Regional)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 7  N = 1
    Change in yearly rate of violent institutional incidents (mean)  -0.089 ns  --
    Change in yearly rate of non-violent institutional incidents (mean)  0.067 ns  --
    Likelihood of discretionary release (Odds Ratio)  --  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 6, X = 1964  N = 1, X = 827
    Difference between time at risk in community  T-statistic = -1.46 ns
    Likelihood of any readmission  0 ns  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Community outcomes where approximated referral criteria met  N = --, X = --  N = --, X = --
    Difference between time at risk in community  T-statistic = --
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = 6, X = 1964  N = 1, X = 8275
    Difference between time at risk in community  T-statistic = -1.46 ns
    Likelihood of any readmission  0 ns  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = --, X = --  N = --, X = --
    Difference between time at risk in community  T-statistic = --
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 Program: LI-SOP (Child) (Local)
 Intent to Treat  Comparison
 Institutional Outcomes  N = 12  N = 5
    Change in yearly rate of violent institutional incidents (mean)  --  --
    Change in yearly rate of non-violent institutional incidents (mean)  -0.042 ns  -1.386 ns
    Likelihood of discretionary release (Odds Ratio)  --  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 15, X = 2241.9  N = 5, X = 2304
    Difference between time at risk in community  T-statistic = 0.14 ns
    Likelihood of any readmission  0.587 ns  1
    Likelihood of readmission with new offence  0 ns  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Community outcomes where approximated referral criteria met  N = 5, X = 2619.8  N = 1, X = 2625
    Difference between time at risk in community  T-statistic = 0.01 ns
    Likelihood of any readmission  0.679 ns  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = 15, X = 2241.9  N = 5, X = 2304
    Difference between time at risk in community  T-statistic = 0.14 ns
    Likelihood of any readmission  0,.587 ns  1
    Likelihood of readmission with new offence  0 ns  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1
    Completions  N = --, X = --  N = --, X = --
    Difference between time at risk in community  T-statistic = --
    Likelihood of any readmission  --  1
    Likelihood of readmission with new offence  --  1
    Likelihood of readmission with new violent offence  --  1
    Likelihood of new sex offence  --  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (non-Aboriginal Male)
 Program: A & EM
 Intent to Treat  Comparison
 Institutional Outcomes  N = 3.803  N = 2.392
    Change in yearly rate of violent institutional incidents (mean)  0.011 ns  -0.060 ns
    Change in yearly rate of non-violent institutional incidents (mean)  0.407***  -0.103 ns
    Likelihood of discretionary release (Odds Ratio)  1.884***  1
    Community Outcomes (ITT): Sample Size (N), Mean time at risk (X)  N = 2,872, X = 1844.9  N = 1,734, X = 1970.2
    Difference between time at risk in community  T-statistic = 5.02***
    Likelihood of any readmission  0.715***  1
    Likelihood of readmission with new offence  0.691***  1
    Likelihood of readmission with new violent offence  0.664***  1
    Completions  N = 2,711, X = 1849.8  N = 1,734, X = 1970.2
    Difference between time at risk in community  T-statistic = 4.78***
    Likelihood of any readmission  0.697***  1
    Likelihood of readmission with new offence  0.665***  1
    Likelihood of readmission with new violent offence  0.676***  1
    Completions  N = 161, X = 1761.9  N = 1,734, X = 1970.2
    Difference between time at risk in community  T-statistic = 3.23***
    Likelihood of any readmission  1.021 ns  1
    Likelihood of readmission with new offence  1.099 ns  1
    Likelihood of readmission with new violent offence  0.406**  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal Male)
 Program: A & EM
 Intent to Treat  Comparison
 Institutional Outcomes  N = 931  N = 687
    Change in yearly rate of violent institutional incidents (mean)  0.059**  -0.112**
    Change in yearly rate of non-violent institutional incidents (mean)  0.490***  0.170 ns
    Likelihood of discretionary release (Odds Ratio)  2.314***  1
    Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 740, X = 2049.1  N = 499, X = 2041.9
    Difference between time at risk in community  T-statistic = -0.16 ns
    Likelihood of any readmission  0.818***  1
    Likelihood of readmission with new offence  0.755***  1
    Likelihood of readmission with new violent offence  0.699 ns  1
    Completions  N = 688, X = 2056.1  N = 499, X = 2041.9
    Difference between time at risk in community  T-statistic = -0.31 ns
    Likelihood of any readmission  0.798***  1
    Likelihood of readmission with new offence  0.726***  1
    Likelihood of readmission with new violent offence  0.667*  1
    Completions  N = 52, X = 1956.8  N = 499, X = 2041.9
    Difference between time at risk in community  T-statistic = 0.79 ns
    Likelihood of any readmission  1,160 ns  1
    Likelihood of readmission with new offence  1,220 ns  1
    Likelihood of readmission with new violent offence  0,973 ns  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Women)
 Program: A & EM
 Intent to Treat  Comparison
 Institutional Outcomes  N = 147  N = 51
    Change in yearly rate of violent institutional incidents (mean)  -0.036 ns  -0.362 **
    Change in yearly rate of non-violent institutional incidents (mean)  -0.244 ns  0.784 ns
    Likelihood of discretionary release (Odds Ratio)  1.276 ns  1
    Community Outcomes (ITT): Sample Size (N), Mean time at risk (X)  N = 145, X = 1905.3  N = 53, X = 1421.6
    Difference between time at risk in community  T-statistic = -4.19***
    Likelihood of any readmission  1.282 ns  1
    Likelihood of readmission with new offence  1.533 ns  1
    Likelihood of readmission with new violent offence  0.414 ns  1
    Completions  N = , X =  N = , X =
    Difference between time at risk in community  T-statistic =
    Likelihood of any readmission    1
    Likelihood of readmission with new offence    1
    Likelihood of readmission with new violent offence    1
    Completions  N = 17, X = 648.12  N = 148, X = 830.93
    Difference between time at risk in community  T-statistic = 3.36***
    Likelihood of any readmission    1
    Likelihood of readmission with new offence    1
    Likelihood of readmission with new violent offence    1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (non-Aboriginal Male)
 Program: R & R
 Intent to Treat  Comparison
 Institutional Outcomes  N = 3,066  N = 1,218
    Change in yearly rate of violent institutional incidents (mean)  0.023*  -0.029 ns
    Change in yearly rate of non-violent institutional incidents (mean)  0.556***  0.043 ns
    Likelihood of discretionary release (Odds Ratio)  1.653**  1
    Community Outcomes (ITT): Sample Size (N), Mean time at risk (X)  N = 2,388, X = 1004.2  N = 807, X = 1206.2
    Difference between time at risk in community  T-statistic = 11.86***
    Likelihood of any readmission  0.729***  1
    Likelihood of readmission with new offence  0.689***  1
    Likelihood of readmission with new violent offence  0.506***  1
    Completions  N = 1,974, X = 1006.6  N = 807, X = 1206.2
    Difference between time at risk in community  T-statistic = 11.21***
    Likelihood of any readmission  0.682***  1
    Likelihood of readmission with new offence  0.638***  1
    Likelihood of readmission with new violent offence  0.413***  1
    Completions  N = 414, X = 992.7  N = 807, X = 1206.2
    Difference between time at risk in community  T-statistic = 7.44***
    Likelihood of any readmission  0.931 ns  1
    Likelihood of readmission with new offence  0.875 ns  1
    Likelihood of readmission with new violent offence  0.963 ns  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal Male)
 Program: R & R
 Intent to Treat  Comparison
 Institutional Outcomes  N = 595  N = 267
    Change in yearly rate of violent institutional incidents (mean)  -0.001 ns  -0.002 ns
    Change in yearly rate of non-violent institutional incidents (mean)  0.419**  0.688 ns
    Likelihood of discretionary release (Odds Ratio)  1.653**  1
    Community Outcomes (ITT): Sample Size (N), Mean time at risk (X)  N = 465, X = 1048.5  N = 178, X = 1185.6
    Difference between time at risk in community  T-statistic = 3.83***
    Likelihood of any readmission  0.807**  1
    Likelihood of readmission with new offence  0.863 ns  1
    Likelihood of readmission with new violent offence  0.755 ns  1
    Completions  N = 370, X = 1046.2  N = 178, X = 1185.6
    Difference between time at risk in community  T-statistic = 3.66***
    Likelihood of any readmission  0.735***  1
    Likelihood of readmission with new offence  0.772 ns  1
    Likelihood of readmission with new violent offence  0.488 ns  1
    Completions  N = 95, X = 1057.4  N = 178, X = 1185.6
    Difference between time at risk in community  T-statistic = 2.23**
    Likelihood of any readmission  1.110 ns  1
    Likelihood of readmission with new offence  1.238 ns  1
    Likelihood of readmission with new violent offence  2.038 ns  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Women)
 Program: R & R
 Intent to Treat  Comparison
 Institutional Outcomes  N = 66  N = 17
    Change in yearly rate of violent institutional incidents (mean)  -0.339*  0.432 ns
    Change in yearly rate of non-violent institutional incidents (mean)  -1.224 ns  0.849 ns
    Likelihood of discretionary release (Odds Ratio)  1.419 ns  1
    Community Outcomes (ITT): Sample Size (N), Mean time at risk (X)  N = 81, X = 1206.9  N = 19, X = 1047.4
    Difference between time at risk in community  T-statistic = -1.65 ns
    Likelihood of any readmission  0.669 ns  1
    Likelihood of readmission with new offence  0.968 ns  1
    Likelihood of readmission with new violent offence  0.001 ns  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Non-Aboriginal)
 Program: Community Maintenance Program
 Intent to Treat  Comparison
 Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 430, X = 1085.6  N = 586, X = 1056.9
    Difference between time at risk in community  T-statistic = -0.76 ns
    Likelihood of any readmission  0.712***  1
    Likelihood of readmission with new offence  0.596***  1
    Likelihood of readmission with new violent offence  0.443***  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 Effectiveness Indicator (Aboriginal)
 Program: Community Maintenance Program
 Intent to Treat  Comparison
 Community Outcomes: Sample Size (N), Mean time at risk (X)  N = 102, X = 897.61  N = 207, X = 907.08
    Difference between time at risk in community  T-statistic = 0.14 ns
    Likelihood of any readmission  0.578***  1
    Likelihood of readmission with new offence  0.465***  1
    Likelihood of readmission with new violent offence  0.414*  1

Note: Hazard Ratios; ns = not significant; *p < .10, **p < .05, ***p < .01


 

Appendix K: Regional Distribution on Those Referred and Enrolled in Correctional Programs

 Program  Region  # Referred  % Enrolled
 Violent Offender
   Pacific  Non-Aboriginal Male  684  36.26
     Aboriginal Male  362  45.58
     Women  50  62.00
   Prairie  Non-Aboriginal Male  245  48.98
     Aboriginal Male  756  61.90
     Women  90  75.56
   Ontario  Non-Aboriginal Male  413  26.88
     Aboriginal Male  75  26.67
   Quebec  Non-Aboriginal Male  362  38.67
     Aboriginal Male  121  37.19
   Atlantic  Non-Aboriginal Male  87  44.83
     Aboriginal Male  11  45.45
 
 Family Violence Prevention
   Pacific  Non-Aboriginal Male  227  25.11
     Aboriginal Male  64  20.31
   Prairie  Non-Aboriginal Male  427  48.01
     Aboriginal Male  514  51.95
   Ontario  Non-Aboriginal Male  887  41.26
     Aboriginal Male  122  34.43
   Quebec  Non-Aboriginal Male  233  41.20
     Aboriginal Male  13  46.15
   Atlantic  Non-Aboriginal Male  383  67.10
     Aboriginal Male  49 61.22
 
 Substance Abuse
   Pacific  Non-Aboriginal Male  986  60.75
     Aboriginal Male  303  49.17
     Women  79  77.22
   Prairie  Non-Aboriginal Male  1,502  67.58
     Aboriginal Male  1,204  57.39
     Women  259  64.48
   Ontario  Non-Aboriginal Male  1,600  46.88
     Aboriginal Male  224  46.88
     Women  168  83.93
   Quebec  Non-Aboriginal Male  920  59.89
    > Aboriginal Male  48  54,17
     Women  69  75.36
   Atlantic  Non-Aboriginal Male  968  72.73
     Aboriginal Male  112  62.50
     Women  104  89.42
 
 Sex Offender
   Pacific  Non-Aboriginal Male  419  51.07
     Aboriginal Male  183  55.74
   Prairie  Non-Aboriginal Male  740  73.38
     Aboriginal Male  603  72.97
   Ontario  Non-Aboriginal Male  1,173  68.54
     Aboriginal Male  159  61.64
   Quebec  Non-Aboriginal Male  1,081  69.66
     Aboriginal Male  93  66.67
   Atlantic  Non-Aboriginal Male  547  84.10
     Aboriginal Male  59  83.05

 

Appendix L: Schedule I Offences

  • Full description of Section 1 Schedule::An offence under any of the following provisions of the Criminal Code, that was prosecuted by way of indictment:
  • (a) section 75 (piratical acts);
  • (a.1) section 76 (hijacking);
  • (a.2) section 77 (endangering safety of aircraft or airport);
  • (a.3) section 78.1 (seizing control of ship or fixed platform);
  • (a.4) paragraph 81(1)(a), (b) or (d) (use of explosives);
  • (a.5) paragraph 81(2)(a) (causing injury with intent);
  • (b) subsection 85(1) (using firearm in commission of offence);
  • (b.1) subsection 85(2) (using imitation firearm in commission of offence);
  • (c) subsection 86(1) (pointing a firearm);
  • (d) section 144 (prison breach);
  • (e) section 151 (sexual interference);
  • (f) section 152 (invitation to sexual touching);
  • (g) section 153 (sexual exploitation);
  • (h) section 155 (incest);
  • (i) section 159 (anal intercourse);
  • (j) section 160 (bestiality, compelling, in presence of or by child);
  • (k) section 170 (parent or guardian procuring sexual activity by child);
  • (l) section 171 (householder permitting sexual activity by or in presence of child);
  • (m) section 172 (corrupting children);
  • (n) subsection 212(2) (living off the avails of prostitution by a child);
  • (o) subsection 212(4) (obtaining sexual services of a child);
  • (o.1) section 220 (causing death by criminal negligence);
  • (o.2) section 221 (causing bodily harm by criminal negligence);
  • (p) section 236 (manslaughter);
  • (q) section 239 (attempt to commit murder);
  • (r) section 244 (discharging firearm with intent);
  • (s) section 246 (overcoming resistance to commission of offence);
  • (s.1) subsections 249(3) and (4) (dangerous operation causing bodily harm and dangerous operation causing death);
  • (s.2) subsections 255(2) and (3) (impaired driving causing bodily harm and impaired driving causing death);
  • (s.3) section 264 (criminal harassment);
  • (t) section 266 (assault);
  • (u) section 267 (assault with a weapon or causing bodily harm);
  • (v) section 268 (aggravated assault);
  • (w) section 269 (unlawfully causing bodily harm);
  • (x) section 270 (assaulting a peace officer);
  • (y) section 271 (sexual assault);
  • (z) section 272 (sexual assault with a weapon, threats to a third party or causing bodily harm);
  • (z.1) section 273 (aggravated sexual assault);
  • (z.2) section 279 (kidnapping);
  • (z.21) section 279.1 (hostage taking);
  • (z.3) section 344 (robbery);
  • (z.31) subsection 430(2) (mischief that causes actual danger to life);
  • (z.32) section 431 (attack on premises, residence or transport of internationally protected person);
  • (z.33) section 431.1 (attack on premises, accommodation or transport of United Nations or associated personnel);
  • (z.34) subsection 431.2(2) (explosive or other lethal device);
  • (z.4) section 433 (arson – disregard for human life);
  • (z.5) section 434.1 (arson – own property);
  • (z.6) section 436 (arson by negligence); and
  • (z.7) paragraph 465(1)(a) (conspiracy to commit murder).
  • Full description of Section 2 Schedule::An offence under any of the following provisions of the Criminal Code, as they read immediately before July 1, 1990, that was prosecuted by way of indictment:


  • (a) section 433 (arson);
  • (b) section 434 (setting fire to other substance); and
  • (c) section 436 (setting fire by negligence).
  • Full description of Section 3 Schedule::An offence under any of the following provisions of the Criminal Code, chapter C-34 of the Revised Statutes of Canada, 1970, as they read immediately before January 4, 1983, that was prosecuted by way of indictment:


  • (a) section 144 (rape);
  • (b) section 145 (attempt to commit rape);
  • (c) section 149 (indecent assault on female);
  • (d) section 156 (indecent assault on male);
  • (e) section 245 (common assault); and
  • (f) section 246 (assault with intent).
  • Full description of Section 4 Schedule::An offence under any of the following provisions of the Criminal Code, chapter C-34 of the Revised Statutes of Canada, 1970, as they read immediately before January 1, 1988, that was prosecuted by way of indictment:

  • (a) section 146 (sexual intercourse with a female under 14);
  • (b) section 151 (seduction of a female between 16 and 18);
  • (c) section 153 (sexual intercourse with step-daughter);
  • (d) section 155 (buggery or bestiality);
  • (e) section 157 (gross indecency);
  • (f) section 166 (parent or guardian procuring defilement); and
  • (g) section 167 (householder permitting defilement).
  • Full description of Section 5 Schedule::The offence of breaking and entering a place and committing an indictable offence therein, as provided for by paragraph 348(1)(b) of the Criminal Code, where the indictable offence is an offence set out in sections 1 to 4 of this Schedule and its commission

  • (a) is specified in the warrant of committal;
  • (b) is specified in the Summons, Information or Indictment on which the conviction has been registered;
  • (c) is found in the reasons for judgment of the trial judge; or
  • (d) is found in a statement of facts admitted into evidence pursuant to section 655 of the Criminal Code.
  • Full description of Section 6 Schedule::An offence under any of the following provisions of the Crimes Against Humanity and War Crimes Act:


  • (a) section 4 (genocide, etc., committed in Canada);
  • (b) section 5 (breach of responsibility committed in Canada by military commanders or other superiors);
  • (c) section 6 (genocide, etc., committed outside Canada); and
  • (d) section 7 (breach of responsibility committed outside Canada by military commanders or other superiors).

 

Appendix M: References

Andrews, D.A. (2001). Principles of effective correctional programs. In L.L. Motiuk & R.C. Serin (Eds.), Compendium 2000 on effective correctional programming (pp. 9-17). Ottawa, ON: Correctional Service of Canada.

Andrews, D.A., & Bonta, J. (2003). The psychology of criminal conduct (4th ed.). Cincinnati, OH: Anderson Publishing.

Andrews, D.A., Dowden, C., & Gendreau, P. (2000). Clinically relevant and psychologically informed approaches to reducing reoffending: A meta-analytic study of human service, risk, need, responsivity and other concerns in justice contexts. Unpublished Manuscript, Carleton University, Ottawa, Ontario, Canada.

Andrews, D.A., Zinger, I., Hoge, R.D., Bonta, J., Gendreau, P., & Cullen, F.T. (1990). Does correctional treatment work? A psychologically informed meta-analysis. Criminology, 28, 369-404.

Annis, H.M., Turner, N.E. & Sklar, S.M. (1997). IDTS Inventory of Drug Taking Situations User’s Guide. Toronto, ON: Addiction Research Foundation.

Anstiss, B. (2003). Just how effective is correctional treatment at reducing re-offending? New Zealand Journal of Psychology, 32, 84-91.

Aos, S., Phipps, P., Barnoski, R., & Lieb, R. (1999). The comparative costs and benefits of programs to reduce crime: A review of national research findings with implications for Washington State. Washington, DC: Washington State Institute for Public Policy.

Aos, S., Phipps, P., Barnoski, R., & Lieb, R. (2001). The comparative costs and benefits of programs to reduce crime. Washington, DC: Washington State Institute for Public Policy.

Bakker, L., Ward, T., Cryer, M., & Hudson, S.M. (1997). Out of the rut: A cognitive-behavioral treatment program for driving-while-disqualified offenders. Behaviour Change, 14, 29-38.

Barratt, E.S. (1994). Impulsivity and aggression in violence and mental disorder: Developments in risk assessment. Monahan and Steadman (eds.). Chicago, IL: The University of Chicago Press.

Beal, C., Weekes, J. R., Millson, W. A., & Eno, J. (1997). Computerized vs. personal interviews for the identification of offender substance abuse problems. Paper presented at the meeting of the Canadian Psychological Association, Toronto, Ontario.

Beck, A.T., Wright, F.D., Newman, C.F. & Liese, B.S. (1993). Cognitive Therapy of Substance Abuse. New York, NY: Guilford Press.

Bell, A., & Flight, J. (in press). An evaluation of the Spirit of a Warrior program for women offenders. Research Report. Ottawa, ON: Research Branch, Correctional Service of Canada.

Berry, S. (1999). The Montgomery House Violence Prevention Programme: An evaluation. Wellington, NZ: New Zealand Department of Corrections.

Blanchette, K. (2001). Effective correctional practice with women offenders. In L.L. Motiuk & R.C. Serin (Eds.), Compendium 2000 on effective correctional programming (pp. 160-173). Ottawa, ON: Correctional Service of Canada.

Blanchette, K. & Moser, A. (2006). Integrated Correctional Intervention Strategies (ICIS) Pilot Project: Phase I Results Report. Ottawa, ON: Reintegration Programs Branch, Correctional Service of Canada.

Boe, R. (2005). Unemployment risk trends and the implications for Canadian federal offenders. Forum on Corrections Research, 17(1), 3-5.

Boe, R., Nafekh, M., Vuong, B., Sinclair, R., Cousineau, C. (2003). The Changing Profile of the Federal Inmate Population: 1997 and 2002 (Research Report R-132).Ottawa, Ontario: Research Branch, Correctional Service Canada.

Boer, D. P., Kroner, D. G., Wong, S., & Cadsky, O. (n.d.). Abusive Relationships Inventory (ARI). Test Battery for the High Intensity Family Violence Prevention Program. Ottawa: Correctional Service of Canada

Bonta, J., Harman, W. G., Hann, R. G., & Cormier, R. B. (1996). The prediction of recidivism among federally sentenced offenders: a re-validation study of the SIR scale. Canadian Journal of Criminology, 38, 61–79.

Brown, S. & Motiuk, L.L. (2005). The Dynamic Factor Identification and Analysis (DFIA) component of the Offender Intake Assessment (OIA) process: A meta-analytic, psychometric and consultative review (Research Report R-164). Ottawa, Ontario: Research Branch, Correctional Service Canada.

Bumby, K. M. (1996). Assessing the cognitive distortions of child molesters and rapists: Development and validation of the MOLEST and RAPE Scales. Sexual Abuse: A journal of research and treatment, 8, 37–54.

Buss, A.H. & Perry, M. (1992). The aggression questionnaire. Journal of personality and social psychology, 63, 452-459.

Cleland, C.M., Pearson, F., & Lipton, D.S. (1996, November). A meta-analytic approach to the link between needs-targeted treatment and reductions in criminal offending. Paper presented at the American Society of Criminology annual meeting. Chicago, IL.

Correctional Service Canada (2003). Commissioner’s Directive CD 726. Correctional Programs. Retrieved from http://infonet/infonet/policies/cd_e.shtml.

Correctional Service Canada (2004). Commissioner’s Directive 720. Education programs and services for offenders. Retrieved from http://infonet/infonet/policies/cd_e.shtml.

Correctional Service of Canada. (2007). 2007-2008 report on plans and priorities. Ottawa, ON: Correctional Service of Canada.

Correctional Service of Canada, Offender Programs and Reintegration Sector (2007). Offender Programs. Unpublished position paper prepared for the Corrections Review Panel.

Correctional Service Canada Reintegration Programs Branch (1993). Revised Anger and Emotions Management Questionnaire. Ottawa, Ontario: Correctional Service Canada.

Correctional Service Canada, Reintegration Programs Branch (2008). Selection Criteria for National Substance Abuse Programs (NSAP). Retrieved from http://infonet/opr/programs/documents/bulletin_15_nsap.doc

Correctional Service Canada, Research Branch (2004). The Changing Federal Offender Population: Profiles and Forecasts. Ottawa, Ontario.

Correctional Service Canada, Research Branch (2006). The Changing Federal Offender Population: Profiles and Forecasts. Ottawa, Ontario.

Correctional Service Canada Review Panel (2007). A Roadmap to Strengthening Public Safety. A Roadmap to Strengthening Public Safety. Retrieved from http://www.publicsafety.gc.ca.

Corrections and Conditional Release Act, R.S.C. 1992, c.20

Cortoni, F., Nunes, K., & Latendresse, M. (2006). An examination of the effectiveness of the violence prevention program. Research Report No. R-178. Ottawa, ON: Research Branch, Correctional Service of Canada.

Donato, R., & Shanahan, M. (1999). The economics of implementing intensive in-prison sex-offender treatment programs. Trends and issues in crime and justice, 134, 1-6.

Donovan, D.M. & O’Leary, M.R. (1978). The Drinking-related locus of control scale: Reliability, factor structure and validity. Journal of studies on alcohol, 39(5), 759-781.

Dowden, C. (1998). A meta-analytic examination of the risk, need and responsivity principles and their importance within the rehabilitation debate. Unpublished Master’s thesis. Carleton University, Ottawa, Ontario, Canada.

Dowden, C., & Andrews, D.A. (1999). What works for female offenders: A meta-analytic review. Crime and Delinquency, 45, 438-452.

Dowden, C., & Andrews, D.A. (2000). Effective correctional treatment and violent re-offending: A meta-analysis. Canadian Journal of Criminology, 42, 449-467.

Dowden, C., Antonowicz, D., & Andrews, D.A. (2003). The effectiveness of relapse prevention with offenders: A meta-analysis. International Journal of Offender Therapy and Comparative Criminology, 47, 516-528.

Dowden, C., Blanchette, K., & Serin, R.C. (1999). Anger management programming for federal male inmates: An effective intervention. Research Report No. R-82. Ottawa, ON: Research Branch, Correctional Service of Canada.

D’Zurilla, T. J., Nezu, A. M., & Maydeu-Olivares, A. (1999). Manual for the Social Problem-Solving Inventory-revised. North Tonawanda, NY: Multi-Health Systems.

Farrington, D.P. & Petrosino, A. (2001). The Campbell Collaboration Crime and Justice Group. The ANNALS of the American Academy of Political and Social Science, 578(1), 35-49.

Fernandez, Y.M., Marshall, W.L., Lightbody, S. & O’ Sullivan, C. (1999). The Child Molester Empathy Measure: Description and examination of its reliability and validity. Sexual Abuse: A Journal of Research and Treatment, 11, 17-31.

Fortin, D. (2004). Program strategy for women offenders. Ottawa, ON: Correctional Service Canada.

French, S.A., & Gendreau, P. (2006). Reducing prison misconducts: What works! Criminal Justice and Behavior, 33, 185-218.

Furlong, A., & Grant, B.A. (2006). Women offender substance abuse programming: Interim results. Forum on Corrections Research, 18, 45-48

Geuens, M., & DePelsmacker P. (2002). Validity and reliability of scores on the reduced Emotional Intensity Scale. Educational and Psychological Measurement. 62, 299-315.

Gibbons, D.C. (1999). Review essay: Changing lawbreakers – What have we learned since the 1950’s? Crime and Delinquency, 45, 272-293.

Gitzel, N. (1997). Family Violence Risk Assessment and Family Violence Programming Prairie Region. Unpublished manuscript.

Grant, B., Kunic, D., MacPerson, P., McKeown, C., & Hanson, E. (2003). The High Intensity Substance Abuse Program (HI-SAP): Results from the pilot programs. Research Report No. R-140. Ottawa, ON: Research Branch, Correctional Service of Canada.

Grant, B., Varis, D.V., & Lefebvre, D. (2004). Intensive support units (ISU) for federal offenders with substance abuse problems: An impact analysis. Research Report No. R-151. Ottawa, ON: Research Branch, Correctional Service of Canada.

Hall, E.A., Prendergast, M.L., Wellisch, J., Patten, M., & Cao, Y. (2004). Treating drug-abusing women prisoners: An outcome evaluation of the forever free program. The Prison Journal, 84, 81-105.

Hanson, R.K., Gordon, A., Harris, A.J.R., Marques, J.K., Murphy, W., Quinsey, V., et al. (2002). First report of the collaborative outcome data project on the effectiveness of psychological treatment for sex offenders. Sexual Abuse: A Journal of Research and Treatment, 14, 169-194.

Hanson, R. K., & Harris, A. J. R. (2001). A structured approach to evaluating change among sexual offenders. Sexual Abuse: A Journal of Research and Treatment, 13, 105-122.

Hanson, R.K., Harris, A.J.R., Scott, T.L. & Helmus, L. (2007). Assessing the risk of sexual offenders on community supervision: The Dynamic Supervision Project. Ottawa, ON: Public Safety Canada.

Hanson, R. K., & Thornton, D. (1999). Static -99: Improving actuarial risk assessments for sex offenders. User Report 99-02. Ottawa: Department of the Solicitor General of Canada. Retrieved from http://www.ps-sp.gc.ca/.

Hill, J.K., Andrews, D.A., & Hoge, R.D. (1991). Meta-analysis of treatment programs for young offenders: The effect of clinically relevant treatment on recidivism, with controls introduced for various methodological variables. Canadian Journal of Program Evaluation, 6, 97-109.

Hupka, R.B. & Rusch, P. (2001). The interpersonal relationship scale. In Perlmutter, B.F., Touliatos, J. & Holden, G.W. (Eds.), Handbook of family measurement techniques: Instruments & index, Vol.3, pp.148-150. Thousand Oaks, CA: Sage Publications

Kennedy, S.M. (2001). Treatment responsivity: Reducing recidivism by enhancing treatment effectiveness. In L.L. Motiuk & R.C. Serin (Eds.), Compendium 2000 on effective correctional programming (pp. 30-36). Ottawa, ON: Correctional Service of Canada.

Kroner, D. G., & Loza, W. (2001). Evidence for the efficacy of self-report in predicting non-violent and violent criminal recidivism. Journal of Interpersonal Violence, 16, 168–177.

Kropp, P.R. & Hart, S.D. (2000). The Spousal Assault Risk Assessment (SARA) guide: Reliability and validity in adult male offenders. Law and human behavior, 24(1), 101-118.

Kropp, P. R., Hart, S. D., Webster, C. D., & Eaves, D. (1995). Manual for the Spousal Assault Risk Assessment Guide (2nd ed.). Vancouver, BC: British Columbia Institute on Family Violence.

Kunic, D. & Grant B. (2006). The Computerized Assessment of Substance Abuse (CASA): Results from the Demonstration Project. CSC: Addictions Research Centre.

Laws, D.R. (1999). Relapse prevention: The state of the art. Journal of Interpersonal Violence, 14, 285-302.

Lightfoot, L. O., & Hodgins, D. (1988). A survey of alcohol and drug problems in incarcerated offenders. International journal of the addictions, 23, 687-706.

Lipsey, M.W., Chapman, G.L., & Landenberger, N.A. (2001). Cognitive-behavioural programs for offenders. Annals of the American Academy of Political and Social Science, 578, 144-157.

Lipsey, M.W., & Cullen, F.T. (in press). The effectiveness of correctional rehabilitation: A review of systematic reviews. Annual Review of Law and Social Science.

Lösel, F., & Schmucker, M. (2005). The effectiveness of treatment for sexual offenders: A comprehensive meta-analysis. Journal of Experimental Criminology, 1, 117-146.

Marshall, W.L. (1992). The social value of treatment for sexual offenders. The Canadian Journal of Human Sexuality, 1, 109-114.

Martin, G.W., Wilkinson, D.A. & Poulos, C. (1995). The Drug Avoidance Self Efficacy Scale. Journal of Substance Abuse, 7(2), 151-163.

McCollister, K.E., French, M.T., Prendergast, M., Wexler, H., Sack, S., & Hall, E. (2003). Is in-prison treatment enough? A cost-effectiveness analysis of prison-based treatment and aftercare services for substance-abusing offenders. Law & Policy, 25, 63-82.

Miller, R.S. & Lefcourt, H.M. (1982). The assessment of social intimacy. Journal of personality assessment, 46(5), 514-518.

Mills, J. F., Kroner, D.G., & Forth, A.E. (2002). Measures of criminal attitudes and associates (MCAA): Development, factor structure, reliability, and validity. Assessment, 9, 240-253.

Mills, J. F., Kroner, D. G., & Hemmati, T (2004) The Measures of Criminal Attitudes and Associates (MCAA): the prediction of general and violent recidivism. Criminal Justice and Behavior 31, 717–733.

Mills, J. F., Loza, W., & Kroner, D. G. (2003) Predictive validity despite social desirability: evidence for the robustness of self-report among offenders. Criminal Behaviour & Mental Health 13, 140–150.

Monahan, J., & Steadman, H. J. (Eds.) (1994). Violence and mental disorder: Developments in risk assessment. Chicago, IL: University of Chicago Press.

Moore, J-P. & Trevethan, S. (2003). First Nations, Métis, Inuit and Non-Aboriginal Federal Offenders: A Comparative Profile (Research Report R-134). Ottawa, Ontario: Research Branch, Correctional Service Canada.

Motiuk, L., Cousineau, C., & Gileno, J. (2005). Safe return of offenders to the community statistical overview. Ottawa, ON: Research Branch, Correctional Service of Canada.

Novaco, R.W. (1994). Novaco Anger Scale and Provocation Inventory (NAS-PI). Los Angeles, CA: Western Psychological Services.

Nufield, J. (1982). Parole decision-making in Canada: research towards decision guidelines. Ottawa, Canada: Ministry of Supply and Services Canada.

Paulhus, D. L. (1984). Two-component models of socially desirable responding. Journal of Personality and Social Psychology, 46, 598-609.

Paulhus, D.L. (1990). Measurement and control response bias. In Robinson, J.P., Shaver, P.R. & Wrightsman, L.S. (Eds.), Measurement of personality and social psychological attitudes. San Diego, CA: Academic Press.

Peacock, E.J. (2000). Warkworth Institution , Correctional Service of Canada

Pearce, S.C., & Holbrook, D. (2002). Research findings and best practices in substance abuse treatment for offenders: A review of the literature. Retrieved August 20, 2007, from http://crrp41.doc.state.nc.us/docs/pubdocs/0006107.PDF.

Pearson, F.S., Lipton, D.S., Cleland, C.M., & Yee, D.S. (2002). The effects of behavioural/ cognitive-behavioural programmes on recidivism. Crime and Delinquency, 48, 476-496.

Peters R.H. (1993). Relapse prevention approaches in the criminal justice system. In T.T. Gorski, J.M. Kelley, L. Havens, & R.H. Peters (Eds.), Relapse prevention and the substance-abusing criminal offender (Technical Assistance Publication Series, No. 8DHHS, Pub. N, SMA, 95-3071). Washington, DC: U.S. Department of Health and Human Services, Center for Substance Abuse Treatment.

Pettway, C. (2006). Best practices tool-kit: Gender-responsive strategies. Ohio Institute on Correctional Best Practices. Retrieved August 20, 2007, from http://www.drc.state.oh.us/web/iej_files/Gender_Responsive_Strategies.pdf.

Polaschek, D.L.L., Wilson, N.J., Townsend, M.R., & Daly, L.R. (2005). Cognitive-behavioural rehabilitation for high-risk violent offenders. Journal of Interpersonal Violence, 20, 1611-1627.

Poporino, F.J., Robinson, D., Millson, W.A., & Weekes, J.R. (2002). An outcome evaluation of prison-based treatment programming for substance abusers. Substance Use and Misuse, 37, 1047-2077.

Posovac, E. J., & Carey, R. G. (1992). Program evaluation: Methods and case studies. Englewood Cliffs, N. J.: Prentice Hall.

Proulx, J., & Perrault, S. (1997). An evaluation of the Ma Mawi Wi Chi Itata family violence program Stony Mountain project. Winnipeg, MB: Ma Mawi Wi Chi Itata.

Robinson, D., Fabiano, E., Porporino, F. P., Millson, W. A., & Graves, G. (1993). A guide to the use of the Computerized Lifestyle Assessment Instrument for substance abuse (2nd ed.). Ottawa, ON: Research and Statistics Branch, Correctional Service Canada.

Robinson, D., & Millson, W. A. (1991). Comparison of computerized interview and file information in the identification of offender substance abuse problems. Unpublished manuscript, Correctional Service of Canada.

Ross, R.R., & Ross, R.D. (Eds.). (1995). Thinking straight: The reasoning and rehabilitation programme for delinquency prevention and offender rehabilitation. Ottawa, ON: AIR Training and Publications.

Serin, R.C., & Preston, D.L. (2001). Violent offender programming. In L.L. Motiuk & R.C. Serin (Eds.), Compendium 2000 on effective correctional programming (pp. 146-157). Ottawa, ON: Correctional Service of Canada.

T3 Associates. (1999). An outcome evaluation of CSC substance abuse programs: OSAPP, ALTO, and Choices: Final report. Ottawa, ON: Correctional Service of Canada

Tellier, C., & Serin, R.C. (2001). The role of staff in effective program delivery. In L.L. Motiuk & R.C. Serin (Eds.), Compendium 2000 on effective correctional programming (pp. 174-184). Ottawa, ON: Correctional Service of Canada.

Tong, L.S.J., & Farrington, D.P. (2006). How effective is the “Reasoning and Rehabilitation” programme in reducing re-offending? A meta-analysis of evaluations in four countries. Psychology, Crime & Law, 12, 3-24.

Trevethan, S., Moore. J.P., & Allegri, N. (2005). The “In Search of Your Warrior” program for Aboriginal offenders: A preliminary evaluation. Research Report No.R-172. Ottawa, ON: Research Branch, Correctional Service of Canada.

Vanderburg, S. A., Weekes, J. R., & Millson, W. A. (1994). Assessment of substance abuse problems in Native offenders: The Computerized Lifestyle Assessment Instrument. Ottawa, ON: Correctional Service Canada.

Ward, T., & Hudson, S.M. (1996). Relapse prevention: A critical analysis. Sexual Abuse: A Journal of Research and Treatment, 8, 177-200.

Weekes, J. R., Moser, A. E., & Langevin, C. M. (1997). Assessing substance abuse offenders for treatment. Workshop presented at the meeting of the International Community Corrections Association, Cleveland, Ohio.

Weekes, J. R., Vanderburg, S. A., & Millson, W. A. (1995). A comparison of the French and English versions of the Computerized Lifestyle Assessment Instrument. Ottawa, Ontario: Correctional Service Canada.

Welsh, B.C. (2004). Monetary costs and benefits of correctional treatment programs: Implications for offender reentry. Federal Probation, 68, 9-13.

Welsh, B.C., & Farrington, D.P. (2000). Correctional intervention programs and cost-benefit analysis. Criminal Justice and Behavior, 27, 115-133.

Wong, S. (2001). Treatment of violent offenders: It works! Paper presented at the Conference of the Swedish Correctional Service. Orebro, Sweden.

Wormith, J. S., & Goldstone, C. S. (1984) The clinical and statistical prediction of recidivism. Criminal Justice and Behavior 11, 3–34.

Yessine, A., & Kroner, D. (2004). Altering antisocial attitudes among federal male offenders on release: A preliminary evaluation of the Counter Point program. Research Report No. R-152. Ottawa, ON: Research Branch, Correctional Service of Canada.

Zellerer, E. (2003). Culturally competent programs: The first family violence program for Aboriginal men in prison. The Prison Journal, 83, 171-190.

 

Footnotes

1 The Correctional Service of Canada (CSC) is the federal government agency responsible for administering sentences imposed by the courts that are two years or more. Sentence administration includes the management of correctional institutions of various security levels and the supervision of offenders under conditional release in the community. More information regarding CSC, including policy and legislation, can be found at http://www.csc-scc.gc.ca.

2 Correctional programs are considered to be distinct from education/employment or social programs. As a result, these programs are not included in this evaluation.

3 A formative evaluation occurs at the mid-point of implementation of a policy, program or initiative, and typically focuses on implementation issues. A summative evaluation incorporates formative evaluation information close to the completion of implementation, and encompasses a broader range of evaluation objectives; namely success, cost-effectiveness, unintended effects and continued relevancy.

4 Correctional Service Canada Review Panel (2007). A Roadmap to Strengthening Public Safety.

5 Sources included but were not limited to: Parole Officers, Program Facilitators, Program Managers, Regional Program Managers, Assistant Wardens Intervention, Managers Assessment and Intervention, Psychologists/Psychotherapists, Parole Officer Supervisors, Elders/Spiritual Advisors, Aboriginal Liaison Officers, Behavioural Counsellors, Program Coordinators, Regional Administrators Correctional Programs, Assistant Deputy Commissioners, Case Management Officers, Clinical Directors/Clinical Team Leaders, Inmate Pay/Employment Officers, Case Management Supervisors, Correctional Manager Psychiatric Hospitals, Personal Development Clerks, Intensive Intervention Strategy Managers, Program Plan Assistants, Guidance Counsellors, and Executive Directors.

6 The evaluation team accepted the written comments of some staff and inmates who did not want to be interviewed but still wanted to provide their opinion, and incorporated these comments by identifying themes commensurate with the qualitative data solicited through the interview keys.

7 A multivariable approach was chosen over a priori subject matching (e.g., propensity score matching, exact matching) as a method for controlling confounding variables and reducing bias, because the former produces similar results and also allows for the examination of the effects of other covariates on the outcome. Also, study participants are less likely to be lost due to an inability to find controls who have the same distribution of matching factors as the study participants (Cepeda, Boston & Strom, 2003; Dohoo, Martin, & Stryhn, 2003; Shah, Laupacis, Hux, & Austin, 2005).

8 A CSC Research report examining this issue is currently being drafted.

9 All Programs includes all items under correctional programs as well as education and personal development programming

10 Correctional programs include all programs addressing criminogenic needs such as community correctional programs, ethno-cultural programs, substance abuse, sex offender programming, living skills, violent offender programs, womens programs, and Aboriginal specified programs.

11 These line items reflects all programs categorized within the group, and only those programs chosen for the evaluation are included in the subsequent break outs, thus, these figures exceed the sum of the selected sub-items below.

12 Non-Aboriginal male offender participants were compared with non-Aboriginal male non-participants, Aboriginal male offender participants with Aboriginal male offender non-participants, and women offender participants with women offender non-participants.

13 CSC is the federal government agency responsible for administering sentences imposed by the courts that are two years or more. Sentence administration includes the management of correctional institutions of various security levels and the supervision of offenders under conditional release in the community. More information regarding CSC, including policy and legislation, can be found at http://www.csc-scc.gc.ca.

14 Correctional programs are considered to be distinct from education/employment or social programs. As a result, these programs are not included in this evaluation.

15 See Appendix A for a description of the Violence Prevention Program, In Search of Your Warrior Program and Spirit of A Warrior Program.

16 See Appendix A for a description of CSCs High and Moderate Intensity Family Violence Prevention programs, and the Aboriginal High Intensity Family Violence Prevention Program.

17 See Appendix A for a description of CSCs High, Moderate and Low Substance Abuse Programs, as well as the Aboriginal Offender Substance Abuse Program and the Women Offender Substance Abuse Program.

18 See Appendix A for a detailed description of CSCs sex offender programs.

19 See Appendix A for a detailed description of CSCs maintenance programs.

20 All Programs includes all items under correctional programs as well as Education and Personal Development Programming

21Correctional programs include all programs addressing criminogenic needs such as community correctional programs, ethno-cultural programs, substance abuse, sex offender programming, living skills, violent offender programs, womens programs, and aboriginal initiatives

22 These line items reflects all programs categorized within the group, and only those programs chosen for the evaluation are included in the subsequent break outs Thus, these figures exceed the sum of the selected sub-items below.

23 See the Measures section for a description of the OIA process.

24 The evaluation team accepted the written comments of some staff and inmates who did not want to be interviewed but still wanted to provide their opinion, and incorporated these comments by identifying themes commensurate with the qualitative data solicited through the interview keys.

25 See Appendix B for the Evaluation Matrix.

26 See Appendix F for key informant interview formats

27 See Appendix L for a listing of Schedule I offences

28 All sex offences are included in Schedule I offences, found in Appendix L

29 Survival analysis is a statistical technique that estimates the time taken to reach some event and the rate of occurrence of that event.

30 Observations are referred to as censored when the dependent variable of interest represents the time to a terminal event (reconviction), and the duration of the study is limited in time.

31 Outcome measures are dependent on resident status (inmate, discretionary release, statutory release with residency), thus a positive parole decision was not examined for those already on parole at the time of the analyses.

32 A multivariable approach was chosen over a priori subject matching (e.g., propensity score matching, exact matching) as a method for controlling confounding variables and reducing bias, because the former produces similar results and also allows for the examination of the effects of other covariates on the outcome. Also, study participants are less likely to be lost due to an inability to find controls who have the same distribution of matching factors as the study participants (Cepeda, Boston & Strom, 2003; Dohoo, Martin, & Stryhn, 2003; Shah, Laupacis, Hux, & Austin, 2005).

33 A research report examining this issue is currently being drafted.

34 A meta-analysis is a method used to summarize effect sizes across independent studies, thereby enabling an overall conclusion to be drawn regarding what works and the factors influencing effectiveness (Lipsey & Wilson, 2001). Unlike narrative reviews (i.e., summaries of outcome studies), systematic reviews employing meta-analytic techniques have specific criteria for including and excluding studies, as well as explicit statements about the methods used to search for studies (Farrington & Petrosino, 2001). As such, meta-analyses are considered to be of superior methodological quality and are relied upon in this review when available.

35 Specifically, studies have reported average reductions in recidivism of 26% to 30% (Andrews et al., 1990; Cleland et al., 1996; Dowden & Andrews, 1999, 2000).

36 See Appendix C for a description of referral criteria.

37 See Appendix D for a description of each program intensity level.

38 SIR-R1 for non-Aboriginal males, and Overall Static Risk Factor levels for Aboriginal males.

39 The AUC (between 0 and 1) measures the probability that non-recidivists would score higher on the actuarial tool of interest than recidivists. An AUC of 1 indicates perfect discrimination between recidivists and non-recidivists, while an AUC of 0.5 or less indicates the scale has no power to discriminate.

40 Although some argue that, since SARA scores are combined according to fixed and explicit algorithms, they can be considered actuarial in nature (Kropp & Hart, 2000). However, the summary ratings of risk derived from these item scores are not made according to fixed and explicit guidelines. As such, the summative results of the SARA are based largely on what is referred to as structured professional judgment. See Appendix G for details regarding the SARA recalibration.

41 The priority rating outcome is based on approximations of the Static 99 and Stable 2000.

42 This sub-sample is composed of those in the larger sample admitted to federal custody between the commencement date of the first VPP session to the end of the study date.

43 Referral criteria include scoring poor or very poor on the SIR, two or more convictions for Schedule I offences and/or homicide, and not meeting referral criteria for a SOP or a high intensity FVPP or SAP. As described in the measures section, whether an offender met the referral criteria was derived from automated data sources such as OMS and CPIC.

44 Note that referral criteria used here were those exclusively for VPP.

45 Because ISOYW was not implemented until 1999, this sample includes those offenders admitted from 2000 until the endpoint of data collection

46 Referral criteria include scoring high risk on the overall static risk level rating, two or more convictions for Schedule I offences and/or homicide, and not meeting referral criteria for a sex offender program or a high intensity family violence or substance abuse program. As described in the measures section, whether an offender met the referral criteria was derived from automated data sources such as OMS and CPIC.

47 Note that referral criteria used here were similar to those of VPP, with overall static risk ratings replacing the SIR-R1 risk rating criteria.

48 Sample was drawn from women offenders in the Pacific and Prairie Regions admitted after SOYW was implemented in 2002.

49 Referral criteria was based on that for the VPP-Women pilot program, and includes two separate incidents of violence (two Schedule I offences) resulting in moderate to high degree of harm, and moderate to high risk to re-offend. This excludes offences against children and sex offences. As described in the measures section, whether an offender met the referral criteria was derived from automated data sources such as OMS and CPIC.

50 Note that referral criteria used here were those exclusively for VPPW.

51 This sub-sample is composed of those in the larger sample admitted to federal custody between the commencement date of the first family violence session to the end of the study date.

52 Referral criteria include scoring poor or very poor on the SIR, two or more convictions for Schedule I offences and/or homicide, and not meeting referral criteria for a SOP or a high intensity FVPP or SAP. As described in the measures section, whether an offender met the referral criteria was derived from automated data sources such as OMS and CPIC.

53 This sub-sample is composed of those in the larger sample admitted to federal custody between the commencement date of the first family violence session to the end of the study date.

54 Referral criteria include scoring poor or very poor on the SIR, two or more convictions for Schedule I offences and/or homicide, and not meeting referral criteria for a SOP or a high intensity FVPP or SAP. As described in the measures section, whether an offender met the referral criteria was derived from automated data sources such as OMS and CPIC.

55 Due to the nature of the data, it had to be assumed that all referrals to substance abuse programs had an established link between crime and substance abuse. As a result, the evaluation team could not assess the number of non-referrals to substance abuse programs who should have been referred based on this criterion.

56 Two-way contingency table analyses revealed that incorporating risk into the referral criteria for this sample did not yield accurate program referrals; specifically, offenders who were referred to high intensity substance abuse programs based on need levels were under-referred to moderate intensity programs when risk was considered in the referral criteria.

57 This sub-sample is composed of those in the larger sample admitted to federal custody between the commencement date of the first NSAP-High session to the end of the study date.

58 This sub-sample is composed of those in the larger sample admitted to federal custody between the commencement date of the first NSAP-High session or AOSAP session to the end of the study date.

59 Offenders are referred to AOSAP if they are assessed as having high or moderate need on the ADS, DAST, and PRD.

60 Women offenders are referred based on a moderate or high substance abuse need. As described in the measures section, whether an offender met the referral criteria was derived from automated data sources such as OMS and CPIC.

61 Due to the nature of the data, it had to be assumed that all referrals to substance abuse programs had an established link between crime and substance abuse. As a result, the evaluation team could not assess the number of non-referrals to substance abuse programs who should have been referred based on this criterion.

62 As described in the measures section, whether an offender met the referral criteria was derived from automated data sources such as OMS and CPIC.

63 This sub-sample is composed of those in the larger sample admitted to federal custody between the commencement date of the first SOP session to the end of the study date.

64 Referral criteria were based on a proxy a measure of the Static 99 and Stable 2007 priority rating (Hanson, Harris, Scott & Helmus, 2007). As described in the measures section, whether an offender met the referral criteria was derived from automated data sources such as OMS and CPIC.

65 This sub-sample is composed of those in the larger sample admitted to federal custody between the commencement date of the first SOP session to the end of the study date.

66 Referral criteria were based on a proxy a measure of the Static 99 and Stable 2007 priority rating (Hanson, Harris, Scott & Helmus, 2007). As described in the measures section, whether an offender met the referral criteria was derived from automated data sources such as OMS and CPIC.

67 VPP has a self-report assessment battery that gauges immediate impacts of program participation through changes in attitudes and beliefs associated with treatment targets such as aggression, impulsivity, and reactions to provocation. The assessment battery contains four measures, three of which are used to gauge treatment performance and one which is used to assess the authenticity of responses. See Appendix H for a detailed description and breakdown of results by assessment battery.

68 Due to a relatively small group of offenders who were waitlisted for this program and failed to receive the programming, an appropriate comparison group had to be created using all available Aboriginal women offenders whether the offender was referred to the program or not. The comparison group did not differ from the study group on any of the following variables: age at intake; risk level; need level; or aggregate sentence length. For this analysis, only the ITT group was appropriate to analyse.

69 Concordance between referral criteria and program assignment for each program is presented in the Implementation section.

70 Concordance between referral criteria and program assignment for each program is presented in the Implementation section.

71 This also includes the Aboriginal High Intensity FVPP.

72 HIFVPP has a self-report assessment battery that gauges immediate impacts of program participation through changes in attitudes and beliefs associated with treatment targets such as jealousy, attitudes and beliefs supporting intimate violence, and programs goals attained through participation, anger management, emotional intensity, reactions to provocation, and impulsivity. The assessment battery contains five self-report measures, four of which are used pre-/post-program to gauge treatment performance and one used to assess the authenticity of responses. For descriptions and detailed results by assessment battery, see Appendix H.

73 See Appendix C for a listing of program referral criteria and a description of how the criteria were approximated for the purposes of this evaluation.

74 Includes High Intensity Aboriginal Offender Substance Abuse Program and the Women Offender Substance Abuse Program.

75 This included the Violence Prevention Program, Aggressive Behaviour Control and In Search of Your Warrior.

76 Includes high intensity AOSAP and WOSAP.

77 Data suggest that living skills programs were typically taken prior to other correctional programs. As such, it was decided to control for the effect of other programs in order to obtain a clearer understanding of the true effect that living skills programs have on the various outcomes. A preliminary analysis revealed that other programs such as violent offender programs were the main contributes of the positive outcomes. Although the analyses which follow do not consider the impact of participation in other correctional programs, they demonstrate that participation in living skills programs provides a clear and significant predictor as to how this group of offenders will perform with regards to institutional and community outcomes.

78 A & EM has a self-report assessment battery that gauges immediate impacts of program participation through changes in attitudes and beliefs associated with treatment targets such as anger management, emotional intensity, reactions to provocation, and impulsivity. The assessment battery contains five self-report measures, four of which are used pre-/post-program to gauge treatment performance and one used to assess the authenticity of responses. See Appendix H for a detailed description and breakdown of results by assessment battery.

79 R & R has a self-report assessment battery that gauges immediate impacts of program participation through changes in attitudes and beliefs associated with treatment targets such as problem solving and impulsivity. The assessment battery contains three measures, two of which are self report and used to gauge treatment performance and 1 completed by the facilitator to assess the offenders motivational level. See Appendix A for details.

80 The dollar return refers to the costs associated with incarcerating an offender minus the costs associated with the offender living in the community.

81 Delivery cost is for the 2006/2007 fiscal year.

82 These line items reflect all programs categorized within the group, and only those programs chosen for the evaluation are included in the subsequent break outs.

83 These line items reflect all programs categorized within the group, and only those programs chosen for the evaluation are included in the subsequent break outs.

84 These line items reflect all programs categorized within the group, and only those programs chosen for the evaluation are included in the subsequent break outs.

85 These line items reflect all programs categorized within the group, and only those programs chosen for the evaluation are included in the subsequent break outs.

86 These line items reflect all programs categorized within the group, and only those programs chosen for the evaluation are included in the subsequent break outs.

87 Note that for all tables in the following analyses, VOP=violent offender programs, and includes ABC, VPP, and ISOYW. FVPP=family violence prevention programs, and includes the National High and Moderate FVPP, as well as the Aboriginal High Intensity FVPP. SAP=substance abuse programs, and includes the National High, Moderate and Low intensity levels, as well as the High and Moderate Intensity Level AOSAP. SOP(a)=sex offender programs where results indicate those for offenders with non-child victims. SOP programs include programs of all intensity levels offered between 1997 and 2007. Numbers in italics indicate the data did not adequately fit the model at the 0.1 level of significance. SOP(b) = sex offender programs where results indicate those for offenders with child victims. SOP programs include programs of all intensity levels offered between 1997 and 2007. Numbers in italics indicate the data did not adequately fit the model at the 0.1 level of significance. Finally LSP=living skills programs, and include A & EM and R & R.

88 Refer to CD 709 on Administrative Segregation for a description of why an offender should be placed in segregation.

89 Five program categories will be reviewed: (a) Substance Abuse; (b) Violent Offender; (c) Family Violence Prevention; (d) Sex Offender; and (e) Maintenance Programs

90 Note that institutional outcomes may vary by program, for example access to private family may be examined as an outcome measure for those with need for offenders who are perpetrators of family violence.

91 All cohorts will be matched on: (a) age groupings; (b) sentence length; (c) risk; (d) motivation level; (e) security level; (f) admission date; (g) need; and (h) program assignment

92 Correctional Service Canada (2003). Standards for Correctional Programs 726-1

93 In-custody snapshot from 1997 to 2002 (Boe et al., 2003)

94 In-custody snapshot in 2002 by Moore & Trevethan (2002)

95 In-custody snapshot from 1997 to 2002 (Boe et al., 2003)

96 Sample based on release cohort (Brown & Motiuk, 2005)